MetPX-Sarracenia Developer’s Guide

version:: UNKNOWN
date:: May 10, 2024

Tools you Need

To hack on the Sarracenia source, you need:

A linux development environment, either a workstation, or a VM of some kind. setup using ubuntu is automated, but adjustment for other distros is possible. command-line comfort a must.
python3. The application is developed in and depends on python versions >= 3.5.
style: PEP8 except max line length is 119, enforced via pycodestyle for VSCode, yapf3 or other similar tool.
an account on github.com will help in submitting patches for consideration.

Things that will be installed by automated setup:

a bunch of other python modules indicated in the dependencies (setup.py or debian/control)
python3 pyftpdlib module, used to run an ftpserver on a high port during the flow test.
git. in order to download the source from the github repository, and to prepare and submit changes.
a dedicated rabbitmq broker, with administrative access, to run the sr_insects tests. this is installed by automated tools for setting up the linux environment. The flow test creates and destroys exchanges and will disrupt any active flows on the broker.

after you have cloned the source code:

git clone -b development https://github.com/MetPX/sarracenia sr3
git clone -b development https://github.com/MetPX/sarrac sr3c
git clone https://github.com/MetPX/sr_insects insects
cd sr3

The rest of the Guide takes the above for granted.

Documentation

Documentation Standards exist in /docs/Contribution/Documentation.rst process for locally building the docs are there, as well as live web-site maintenance methods.

Where to Put Options

Options are documented in sr3_options(7) dictionary style in alphabetic order. Should it be worthwhile, examples of use could be added to other guides.

Development

In general, the development workflow is to get a laptop or a VM where one can run the flow_tests (available from http://github.com/MetPX/sr_insects ) The first step in configuring a development environment is ensuring that the sr_insects flow tests work, as they function as a gate for commits to important branches.

Development is most commonly done on Ubuntu >=18.04 platform.

v2 Workflow

Finished development work for version 2 is committed to on the v2_dev branch, which is used to produce daily snapshots. One should not normally commit changes to the v2_dev branch, but rather merge them from a working branch.

Development branches are named after the issue they are meant to address “v2_issue365”, for example. If there are multiple attempts to address a given issue, then use the issue as a name prefix. For example, there could be issue365, but if we decide that isn’t a good way to address the issue, there could be an issue365_methodB branch.

Before submitting a pull-request (PR), please ensure that the flow tests from sr_insects have been run successfully: at least static_flow, flakey_broker, and dynamic_flow

When a PR is generated, the second developer can look it over for concerns. Once satisfied with the nature of the patch, the second developer should pull the branch and run the flow tests again (the same three) to confirm. Only after the flow tests have been run on multiple machines should a change be merged to stable.

issues unique to v2 should be tagged v2only. on Launchpad.net:

daily repository packages of v2 will be build from v2_dev

pre-release repository packages of v2 will be build from v2_dev

release repository packages are generated from v2_stable.

v3 Workflow

The upcoming version of Sarracenia is developed in the development (work in progress) branch. As the major refactor is substantially complete, the remaining work is now entirely constructive and all development is co-ordinated through issues exactly as v2 is. Issues unique to v3, be they regressions or enhancements that don’t make sense to add to v2, have the tag v3only. Issues that are common between the releases are tagged v3.

The workflow with v3 is similar to v2 but with different branches. branches are assumed to be branched from the development branch, so v3 is assumed unless v2_ is present. Having all the flow tests complete fairly successfully is one criterion for acceptance into development.

To run the sr_insects tests, the repository must be cloned with the development branch. A gate for merging to development is for a second developer to run the flow_tests. For v03, these tests must run: static_flow, flakey_broker, dynamic_flow, transform_flow

launchpad has recipes to produce metpx-sr3 packages from various branches.

The MetPX Daily repository is a snapshot of the development branch.

The MetPX Pre-Release repository should receive versions ending in rcX (release candidate) The packages here from from pre-release branch which comes from snapshots of the development branch. There is also a pre-release-py36 branch for building pre-release packages for older operating systems.

stable comes from on snapshots of (version 3) pre-release branch.

The MetPX repository should only contain stable releases that have graduated from the rcX series. there is a stable_py36 branch to build packages for older operating systems that have python 3.6 (redhat 8, ubuntu 18, ubuntu 20) or are too old to use hatchling installer.

sr_insects

The sr_insects repository has it’s own issues DB, and work on sr_insects is encouraged. Both v2 and v3 are supported on the stable branch of sr_insects. That branch should be used to support all development in both versions….

Local Installation

There are many different ways to install python packages on a computer. Different developers will prefer different methods, and all the methods need to be tested prior to each release. Sarracenia can work with either mqtt or amqp (most mature and stable) message passing libraries. Install one of those first. in these examples, we use amqp.

Wheel when people are running different operating systems (non-ubuntu, non-debian) people will be installing wheels, typically that have been uploaded to pypi.python.org. On the other hand, it is a bit of a pain/noise to upload every development version, so we only upload releases, so testing of wheels is done by building local wheels. Need to build a new wheel every time a change is made.
pip install metpx-sr3[amqp] would pull a wheel down from pypi.python.org. Generally not used during development of Sarracenia itself. one could also pull in all possible dependencies with pip install metpx-sr3[all]
pip install -e .[amqp] … lets you edit the source code of the installed package, ideal for debugging problems, because it allows live changes to the application without having to go through building and installing a new package.
apt install metpx-sr3 install debian package from repositories, similarly to pip install (not -e), normally dev snapshots are not uploaded to repositories, so while this would be the normal way for users of ubuntu servers, it is not available during development of the package itself. Also need apt install python3-amqp
dpkg -i builds a debian package for local installation. This is how packages are tested prior to upload to repositories. It can also be used to support development (have to run dpkg -i for each package change.) also need apt install python3-amqp

The sr_insects tests invokes the version of metpx-sarracenia that is installed on the system, and not what is in the development tree. It is necessary to install the package on the system in order to have it run the sr_insects tests.

Prepare a Vanilla VM

This section describes creating a test environment for use in a virtual machine. One way to build a virtual machine is to use multipass (https://multipass.run) Assuming it is installed, one can create a vm with:

multipass launch -m 8G -d 30G --name flow

need to have ssh localhost work in the multipass container. Can do that by copying multipass private key into the container:

fractal% multipass list
Name                    State             IPv4             Image
primary                 Stopped           --               Ubuntu 20.04 LTS
flow                    Running           10.23.119.56     Ubuntu 20.04 LTS
keen-crow               Running           10.23.119.5      Ubuntu 20.04 LTS
fractal%

Weird issues with ssh keys not being interpreted properly by paramiko, work around ( https://stackoverflow.com/questions/54612609/paramiko-not-a-valid-rsa-private-key-file )

fractal% sudo cat /var/snap/multipass/common/data/multipassd/ssh-keys/id_rsa | sed 's/BEGIN .*PRIVATE/BEGIN RSA PRIVATE/;s/END .*PRIVATE/END RSA PRIVATE/' >id_rsa_container
chmod 600 id_rsa_container
scp -i id_rsa_container id_rsa_container ubuntu@10.23.119.175:/home/ubuntu/.ssh/id_rsa
                                                                  100% 1704     2.7MB/s   00:00

fractal% scp -i id_rsa_container id_rsa_container ubuntu@10.23.119.106:/home/ubuntu/.ssh/id_rsa
The authenticity of host '10.23.119.106 (10.23.119.106)' can't be established.
ECDSA key fingerprint is SHA256:jlRnxV7udiCBdAzCvOVgTu0MYJR5+kYzNwy/DIhkeD8.
Are you sure you want to continue connecting (yes/no/[fingerprint])? yes
Warning: Permanently added '10.23.119.106' (ECDSA) to the list of known hosts.
id_rsa_container                                                                                                                         100% 1712     9.4MB/s   00:00
fractal% multipass shell flow
Welcome to Ubuntu 20.04.3 LTS (GNU/Linux 5.4.0-81-generic x86_64)

 * Documentation:  https://help.ubuntu.com
 * Management:     https://landscape.canonical.com
 * Support:        https://ubuntu.com/advantage

  System information as of Fri Aug 27 21:12:16 EDT 2021

  System load:  0.42              Processes:             112
  Usage of /:   4.4% of 28.90GB   Users logged in:       0
  Memory usage: 5%                IPv4 address for ens4: 10.23.119.106
  Swap usage:   0%


1 update can be applied immediately.
To see these additional updates run: apt list --upgradable


To run a command as administrator (user "root"), use "sudo <command>".
See "man sudo_root" for details.

ubuntu@flow:~$

then prompt ssh to accept the localhost key:

ubuntu@flow:~$ ssh localhost ls -a
The authenticity of host 'localhost (127.0.0.1)' can't be established.
ECDSA key fingerprint is SHA256:jlRnxV7udiCBdAzCvOVgTu0MYJR5+kYzNwy/DIhkeD8.
Are you sure you want to continue connecting (yes/no/[fingerprint])? yes
Warning: Permanently added 'localhost' (ECDSA) to the list of known hosts.
.
..
.bash_logout
.bashrc
.cache
.profile
.ssh
ubuntu@flow:~$

This will provide a shell in an initialized VM. To configure it:

git clone -b development https://github.com/MetPX/sarracenia sr3
cd sr3

There are scripts that automate the installation of necessary environment to be able to run tests:

travis/flow_autoconfig.sh
travis/add_sr3.sh

You should be able to see an empty configuration:

sr3 status

sr3c and sr3 are now installed, and should be ready to run a flow test from the sr_insects module, which has also been cloned:

cd ../sr_insects

The v03 branch of sr_insects supports testing of both v2 and v3, and both versions are now installed. The flow tests are intended to be run to confirm compatibility between v2 and v3, and so one must be able to test v2 as well:

ubuntu@flow:~/sr_insects$ dpkg -l | grep metpx
ii  metpx-libsr3c                    3.21.08a1-0~202108270410~ubuntu20.04.1 amd64        C-Implementation of a Sarracenia Client
ii  metpx-sarracenia                 2.21.08-0~202108241854~ubuntu20.04.1   all          Directory mirroring in real-time for users, file servers and web sites.
ii  metpx-sr3                        3.00.008exp                            all          v3 Directory mirroring in real-time for users, file servers and web sites.
ii  metpx-sr3c                       3.21.08a1-0~202108270410~ubuntu20.04.1 amd64        C-Implementation of a Sarracenia Client
ubuntu@flow:~/sr_insects$

The v2 package is metpx-sarracenia, whereas the v3 one is metpx-sr3. the flow tests will detect which version is installed and test v3 if both are present. To override that:

ubuntu@flow:~/sr_insects$ export sarra_py_version=2.21.08
ubuntu@flow:~/sr_insects$

Then one can run flow_tests from this shell normally.

Ubuntu 18.04

A number of systems run Ubuntu 18.04 even though it is pretty old.

‘’’

multipass launch -m 8G bionic

‘’’

can run developer tests as per multipass as described above.

Python Wheel

If you have not used add_sr3.sh (which builds a debian package), then one can use this procedure for local installation on a computer with a python wheel for testing and development:

python3 setup.py bdist_wheel

or… on newer systems, using build instead:

python3 -m build --no-isolation

Should build a wheel in the dist sub-directory. then as root install that new package:

pip3 install --upgrade ...<path>/dist/metpx*.whl

Local Pip install

For local installation on a computer, using a pip For testing and development:

pip3 install -e .
export PATH=${HOME}/.local/bin:${PATH}

Using the local python package installer (PIP) to create a locally editable version. The above will install the package in ~/.local/bin… so need to ensure the path includes that directory.

Debian/Ubuntu

For local installation on a computer, using a debian package. This process builds a local .deb in the parent directory using standard debian mechanisms. - Check the build-depends line in debian/control for dependencies that might be needed to build from source. - The following steps will build sarracenia but not sign the changes or the source package:

cd metpx/sarracenia
sudo apt-get install devscripts
debuild -uc -us
sudo dpkg -i ../<the package just built>

which accomplishes the same thing using debian packaging. The options are detailed below:

Committing Code

What should be done prior to committing to the development branch? Checklist:

do development on some other branch. Usually the branch will be named after the issue being addressed. Example: issue240, if we give up on an initial approach and start another one, there may be issue240_2 for a second attempt. There may also be feature branches, such as v03.
sr_insects tests works (See Testing) The development branch should always be functional, do not commit code if the sr_insects tests are not working.
Natural consequence: if the code changes means tests need to change, include the test change in the commit.
update doc/ manual pages should get their updates ideally at the same time as the code.

Usually there will be many such cycles on an issueXXX branch before one is ready to issue a pull request. Eventually, we get to Commits to the Development Branch

sr_insects Tests Description

Before committing code to the stable branch, as a Quality Assurance measure, one should run all available self-tests. It is assumed that the specific changes in the code have already been unit tested. Please add self-tests as appropriate to this process to reflect the new ones. Generally speaking one should solve problems at the first test that fails as each test is more complicated than the previous one.

There is a separate git repository containing the more complex tests https://github.com/MetPX/sr_insects

A typical development workflow will be (Do not try this, this is just an overview of the steps that will be explained in detail in following sections):

git branch issueXXX
git checkout issueXXX
cd sarra ; *make coding changes*
cd ..
debuild -uc -us
cd ../sarrac
debuild -uc -us
sudo dpkg -i ../*.deb
cd ..

git clone -b development https://github.com/MetPX/sr_insects
cd sr_insects
sr3 status  # make sure there are no components configured before you start.
            # test results will likely be skewed otherwise.
for test in unit static_flow flakey_browser transform_flow dynamic_flow; do
   cd $test
   ./flow_setup.sh  # *starts the flows*
   ./flow_limit.sh  # *stops the flows after some period (default: 1000) *
   ./flow_check.sh  # *checks the flows*
   ./flow_cleanup.sh  # *cleans up the flows*
   cd ..
done

#assuming all the tests pass.
git commit -a  # on the branch...

Unit

The unit test in sr_insects is the shortest one taking a minute or so, and not requiring much configuration at all. They are sanity tests of code behaviour. Generally takes a minute or two on a laptop.

Static Flow

The static_flow tests are a bit more complicated, testing more components, using single threaded components in a linear way (all data moves uniformly forward.) It should be more straight-forward to identify issues as there is no deletion and so it lends itself well to repeating subset tests to identify individual issues. It takes about two minutes on a laptop.

Flakey Broker

The flakey_broker tests are the same as the static_flow, but slowed down so that they last a few minutes, and the broker is shutdown and restarted while the posting is happenning. Note that post_log prints before a notification message is posted (because post_log is an on_post plugin, and that action, allows one to modify the notification message, so it needs to be before the post actually happens.)

Dynamic Flow

The dynamic_flow test add advanced features: multi-instances, the winnow component, retry logic testing, and includes file removals as well. Most of the documentation here refers to runnig the dynamic_flow test, as it is the most complicated one, and the ancestor of the others. The unit test was separated out from the beginnig of the dynamic_flow test, and the static_flow is a simplified version of the original flow test as well.

Generally speaking, one should run the tests in sequence and ensure the results of earlier tests are good before proceeding to the next test.

Note that the development system must be configured for the sr_insects tests to run successfully. See the next section for configuration instructions. For development with a fresh OS installation, the configuration steps have been automated and can be applied with the flow_autoconfig.sh script in sr_insects (https://github.com/MetPX/sr_insects/blob/stable/flow_autoconfig.sh). Blind execution of this script on a working system may lead to undesirable side effects; you have been warned!

The configuration one is trying to replicate:

Following table describes what each element of the dynamic flow test does, and the test coverage shows functionality covered.

Configuration	Does	Test Coverage
subscribe t_ddx	copy from data mart to local broker posting notification messages to local xwinno00 and xwinnow01 exchanges.	read amqps public data mart (v02) as ordinary user. shared queue and multiple processes 3 instances download from each q post amqp to a local exchange (v02) as feeder(admin) user post_exchangeSplit to xwinnow0x
winnow t0x_f10	winnow processing publish for xsarra exchange for downloading. as two sources identical, only half notification messages are posted to next	read local amqp v02 as feeder user. complete caching (winnow) function post amqp v02 to local excchange.
sarra download f20	download the winnowed data from the data mart to a local directory (TESTDOCROOT= ~/sarra_devdocroot) add a header at application layer longer than 255 characters.	read local amqp v02 (xsarra) download using built-in python shared queue and multiple processes 5 instances download from each q download using accel_wget plugin AMQP header truncation on publish. post amqp v02 to xpublic as feeder user as http downloads from localhost
subscribe t	download as client from localhost to downloaded_by_sub_t directory.	read amqp from local broker as ordinary user/client. shared queue and multiple processes 5 instances download from each q
watch f40	watch downloaded_by_sub_t (post each file that appears there.) memory ceiling set low	client v03 post of local file. (file: url) auto restarting on memory ceiling.
sender tsource2send	read local file, send via sftp to sent_by_tsource2send directory post to xs_tsource_output	client consume v03 notification messages. consumer read local file. send via sftp. plugin replace_dir posting sftp url. post v02 (converting v03 back.) test post_exchangeSuffix option.
subscribe u_sftp_f60	download via sftp from localhost putting files in downloaded_by_sub_u directory.	client sftp download. accel_sftp plugin.
post test2_f61	post files in sent_by_tsource2send with ftp URL’s in the xs_tsource_poll exchange (wrapper script calls post)	explicit file posting ftp URL posting. post_exchangeSuffix option
poll f62	poll sent_by_tsource2send directory posting sftp download URL’s	polling post_exchangeSuffix option
subscribe ftp_f70	subscribe to test2_f61 ftp’ posts. download files from localhost to downloaded_by_sub_u directory.	ftp url downloading.
subscribe q_f71	subscribe to poll, downloading to recd_by_srpoll_test1	confirming poll post quality.
shovel pclean f90	clean up files so they don’t accumulate fakes failures to exercise retries	shovel function. retry logic.
shovel pclean f91	clean up files so they don’t accumulate	shovel with posting v03 retry logic.
shovel pclean f92	clean up files so they don’t accumulate	shovel with consuming v03 posting v02. retry logic.

Assumption: test environment is a Linux PC, either a laptop/desktop, or a server on which one can start a browser. If working with the C implementation as well, there are also the following flows defined:

Running Flow Tests

This section documents these steps in much more detail. Before one can run the sr_insects tests, some pre-requisites must be taken care of. Note that there is Github Actions integration for at least the development branch to verify functionality on a variety of python version. Consult:

https://github.com/MetPX/sarracenia/actions

Note

for the latest test results. Note that the results include dozens of tests, and are a bit unreliable, typically it may take a few retries for it to work completely (3 or 4 fail after initial attempt, then re-run the failed ones, and then perhaps 1 or two will be left, and on the third pass the last one passes.)

Install Servers on Workstation

To prepare a computer to run the flow test, one must install some server software and configurations. This same work is done by travis/flow_autoconfig.sh which is run in Prepare a Vanilla VM but if you need to configure it manually, below is the process.

Install a minimal localhost broker and configure rabbitmq test users:

sudo apt-get install rabbitmq-server
sudo rabbitmq-plugins enable rabbitmq_management

mkdir ~/.config/sarra
cat > ~/.config/sarra/default.conf << EOF
declare env FLOWBROKER=localhost
declare env MQP=amqp
declare env SFTPUSER=${USER}
declare env TESTDOCROOT=${HOME}/sarra_devdocroot
declare env SR_CONFIG_EXAMPLES=${HOME}/git/sarracenia/sarra/examples
EOF

RABBITMQ_PASS=S0M3R4nD0MP4sS
cat > ~/.config/sarra/credentials.conf << EOF
amqp://bunnymaster:${RABBITMQ_PASS}@localhost/
amqp://tsource:${RABBITMQ_PASS}@localhost/
amqp://tsub:${RABBITMQ_PASS}@localhost/
amqp://tfeed:${RABBITMQ_PASS}@localhost/
amqp://anonymous:${RABBITMQ_PASS}@localhost/
amqps://anonymous:anonymous@hpfx.collab.science.gc.ca
amqps://anonymous:anonymous@hpfx1.collab.science.gc.ca
amqps://anonymous:anonymous@hpfx2.collab.science.gc.ca
amqps://anonymous:anonymous@dd.weather.gc.ca
amqps://anonymous:anonymous@dd1.weather.gc.ca
amqps://anonymous:anonymous@dd2.weather.gc.ca
ftp://anonymous:anonymous@localhost:2121/
EOF

cat > ~/.config/sarra/admin.conf << EOF
cluster localhost
admin amqp://bunnymaster@localhost/
feeder amqp://tfeed@localhost/
declare source tsource
declare subscriber tsub
declare subscriber anonymous
EOF

sudo rabbitmqctl delete_user guest

sudo rabbitmqctl add_user bunnymaster ${RABBITMQ_PASS}
sudo rabbitmqctl set_permissions bunnymaster ".*" ".*" ".*"
sudo rabbitmqctl set_user_tags bunnymaster administrator

sudo systemctl restart rabbitmq-server
cd /usr/local/bin
sudo mv rabbitmqadmin rabbitmqadmin.1
sudo wget http://localhost:15672/cli/rabbitmqadmin
sudo chmod 755 rabbitmqadmin

sr3 --users declare

Note

Please use other passwords in credentials for your configuration, just in case. Passwords are not to be hard coded in self test suite. The users bunnymaster, tsource, tsub, and tfeed are to be used for running tests.

The idea here is to use tsource, tsub, and tfeed as broker accounts for all self-test operations, and store the credentials in the normal credentials.conf file. No passwords or key files should be stored in the source tree, as part of a self-test suite.

Setup Flow Test Environment

Once the server environment is established, the flow tests use sftp transfers to localhost.

It is also required that passwordless ssh access is configured on the test host for the system user that will run the flow test. This can be done by creating a private/public ssh key pair for the user (if there isn’t one already) and copying the public key to the authorized_keys file in the same directory as the keys (~/.ssh). For associated commands, see http://www.linuxproblem.org/art_9.html

Note that on systems where older versions of Paramiko (< 2.7.2) are installed, and the ssh key pair was generated with OpenSSH >= 6.5, manually testing the below command will work, but Paramiko will not be able to connect. This is likely the case if the ~/.ssh/id_rsa file contains BEGIN OPENSSH PRIVATE KEY. To work around this, convert the private key’s format using ssh-keygen -p -m PEM -f ~/.ssh/id_rsa.

To confirm that that passwordless ssh to localhost works:

ssh localhost ls

This should run and complete. If it prompts for a password, the flow tests will not work.

Check that the broker is working:

systemctl status rabbitmq-server

One part of the flow test runs an sftp server, and uses sftp client functions. Need the following package for that:

sudo apt-get install python3-pyftpdlib python3-paramiko

The setup script starts a trivial web server, and ftp server, and a daemon that invokes sr_post. It also tests the C components, which need to have been already installed as well and defines some fixed test clients that will be used during self-tests:

cd
git clone https://github.com/MetPX/sr_insects
cd sr_insects
cd static_flow
. ./flow_setup.sh

blacklab% ./flow_setup.sh
cleaning logs, just in case
rm: cannot remove '/home/peter/.cache/sarra/log/*': No such file or directory
Adding flow test configurations...
2018-02-10 14:22:58,944 [INFO] copying /usr/lib/python3/dist-packages/sarra/examples/cpump/cno_trouble_f00.inc to /home/peter/.config/sarra/cpump/cno_trouble_f00.inc.
2018-02-10 09:22:59,204 [INFO] copying /home/peter/src/sarracenia/sarra/examples/shovel/no_trouble_f00.inc to /home/peter/.config/sarra/shovel/no_trouble_f00.inc
2018-02-10 14:22:59,206 [INFO] copying /usr/lib/python3/dist-packages/sarra/examples/cpost/veille_f34.conf to /home/peter/.config/sarra/cpost/veille_f34.conf.
2018-02-10 14:22:59,207 [INFO] copying /usr/lib/python3/dist-packages/sarra/examples/cpump/pelle_dd1_f04.conf to /home/peter/.config/sarra/cpump/pelle_dd1_f04.conf.
2018-02-10 14:22:59,208 [INFO] copying /usr/lib/python3/dist-packages/sarra/examples/cpump/pelle_dd2_f05.conf to /home/peter/.config/sarra/cpump/pelle_dd2_f05.conf.
2018-02-10 14:22:59,208 [INFO] copying /usr/lib/python3/dist-packages/sarra/examples/cpump/xvan_f14.conf to /home/peter/.config/sarra/cpump/xvan_f14.conf.
2018-02-10 14:22:59,209 [INFO] copying /usr/lib/python3/dist-packages/sarra/examples/cpump/xvan_f15.conf to /home/peter/.config/sarra/cpump/xvan_f15.conf.
2018-02-10 09:22:59,483 [INFO] copying /home/peter/src/sarracenia/sarra/examples/poll/f62.conf to /home/peter/.config/sarra/poll/f62.conf
2018-02-10 09:22:59,756 [INFO] copying /home/peter/src/sarracenia/sarra/examples/post/shim_f63.conf to /home/peter/.config/sarra/post/shim_f63.conf
2018-02-10 09:23:00,030 [INFO] copying /home/peter/src/sarracenia/sarra/examples/post/test2_f61.conf to /home/peter/.config/sarra/post/test2_f61.conf
2018-02-10 09:23:00,299 [INFO] copying /home/peter/src/sarracenia/sarra/examples/report/tsarra_f20.conf to /home/peter/.config/sarra/report/tsarra_f20.conf
2018-02-10 09:23:00,561 [INFO] copying /home/peter/src/sarracenia/sarra/examples/report/twinnow00_f10.conf to /home/peter/.config/sarra/report/twinnow00_f10.conf
2018-02-10 09:23:00,824 [INFO] copying /home/peter/src/sarracenia/sarra/examples/report/twinnow01_f10.conf to /home/peter/.config/sarra/report/twinnow01_f10.conf
2018-02-10 09:23:01,086 [INFO] copying /home/peter/src/sarracenia/sarra/examples/sarra/download_f20.conf to /home/peter/.config/sarra/sarra/download_f20.conf
2018-02-10 09:23:01,350 [INFO] copying /home/peter/src/sarracenia/sarra/examples/sender/tsource2send_f50.conf to /home/peter/.config/sarra/sender/tsource2send_f50.conf
2018-02-10 09:23:01,615 [INFO] copying /home/peter/src/sarracenia/sarra/examples/shovel/t_dd1_f00.conf to /home/peter/.config/sarra/shovel/t_dd1_f00.conf
2018-02-10 09:23:01,877 [INFO] copying /home/peter/src/sarracenia/sarra/examples/shovel/t_dd2_f00.conf to /home/peter/.config/sarra/shovel/t_dd2_f00.conf
2018-02-10 09:23:02,137 [INFO] copying /home/peter/src/sarracenia/sarra/examples/subscribe/cclean_f91.conf to /home/peter/.config/sarra/subscribe/cclean_f91.conf
2018-02-10 09:23:02,400 [INFO] copying /home/peter/src/sarracenia/sarra/examples/subscribe/cdnld_f21.conf to /home/peter/.config/sarra/subscribe/cdnld_f21.conf
2018-02-10 09:23:02,658 [INFO] copying /home/peter/src/sarracenia/sarra/examples/subscribe/cfile_f44.conf to /home/peter/.config/sarra/subscribe/cfile_f44.conf
2018-02-10 09:23:02,921 [INFO] copying /home/peter/src/sarracenia/sarra/examples/subscribe/clean_f90.conf to /home/peter/.config/sarra/subscribe/clean_f90.conf
2018-02-10 09:23:03,185 [INFO] copying /home/peter/src/sarracenia/sarra/examples/subscribe/cp_f61.conf to /home/peter/.config/sarra/subscribe/cp_f61.conf
2018-02-10 09:23:03,455 [INFO] copying /home/peter/src/sarracenia/sarra/examples/subscribe/ftp_f70.conf to /home/peter/.config/sarra/subscribe/ftp_f70.conf
2018-02-10 09:23:03,715 [INFO] copying /home/peter/src/sarracenia/sarra/examples/subscribe/q_f71.conf to /home/peter/.config/sarra/subscribe/q_f71.conf
2018-02-10 09:23:03,978 [INFO] copying /home/peter/src/sarracenia/sarra/examples/subscribe/t_f30.conf to /home/peter/.config/sarra/subscribe/t_f30.conf
2018-02-10 09:23:04,237 [INFO] copying /home/peter/src/sarracenia/sarra/examples/subscribe/u_sftp_f60.conf to /home/peter/.config/sarra/subscribe/u_sftp_f60.conf
2018-02-10 09:23:04,504 [INFO] copying /home/peter/src/sarracenia/sarra/examples/watch/f40.conf to /home/peter/.config/sarra/watch/f40.conf
2018-02-10 09:23:04,764 [INFO] copying /home/peter/src/sarracenia/sarra/examples/winnow/t00_f10.conf to /home/peter/.config/sarra/winnow/t00_f10.conf
2018-02-10 09:23:05,027 [INFO] copying /home/peter/src/sarracenia/sarra/examples/winnow/t01_f10.conf to /home/peter/.config/sarra/winnow/t01_f10.conf
Initializing with sr_audit... takes a minute or two
OK, as expected 18 queues existing after 1st audit
OK, as expected 31 exchanges for flow test created.
Starting trivial http server on: /home/peter/sarra_devdocroot, saving pid in .httpserverpid
Starting trivial ftp server on: /home/peter/sarra_devdocroot, saving pid in .ftpserverpid
running self test ... takes a minute or two
sr_util.py TEST PASSED
sr_credentials.py TEST PASSED
sr_config.py TEST PASSED
sr_cache.py TEST PASSED
sr_retry.py TEST PASSED
sr_consumer.py TEST PASSED
sr_http.py TEST PASSED
sftp testing start...
sftp testing config read...
sftp testing fake message built ...
sftp sr_ftp instantiated ...
sftp sr_ftp connected ...
sftp sr_ftp mkdir ...
test 01: directory creation succeeded
test 02: file upload succeeded
test 03: file rename succeeded
test 04: getting a part succeeded
test 05: download succeeded
test 06: onfly_checksum succeeded
Sent: bbb  into tztz/ddd 0-5
test 07: download succeeded
test 08: delete succeeded
Sent: bbb  into tztz/ddd 0-5
Sent: bbb  into tztz/ddd 0-5
Sent: bbb  into tztz/ddd 0-5
Sent: bbb  into tztz/ddd 0-5
Sent: bbb  into tztz/ddd 0-5
/home/peter
/home/peter
test 09: bad part succeeded
sr_sftp.py TEST PASSED
sr_instances.py TEST PASSED
OK, as expected 9 tests passed
Starting flow_post on: /home/peter/sarra_devdocroot, saving pid in .flowpostpid
Starting up all components (sr start)...
done.
OK: sr3 start was successful
Overall PASSED 4/4 checks passed!
blacklab%

As it runs the setup, it also executes all existing unit_tests. Only proceed to the flow_check tests if all the tests in flow_setup.sh pass.

Run A Flow Test

The flow_check.sh script reads the log files of all the components started, and compares the number of notification messages, looking for a correspondence within +- 10% It takes a few minutes for the configuration to run before there is enough data to do the proper measurements:

./flow_limit.sh

sample output:

initial sample building sample size 8 need at least 1000
sample now   1021
Sufficient!
stopping shovels and waiting...
2017-10-28 00:37:02,422 [INFO] sr_shovel t_dd1_f00 0001 stopping
2017-10-28 04:37:02,435 [INFO] 2017-10-28 04:37:02,435 [INFO] info: instances option not implemented, ignored.
info: instances option not implemented, ignored.
2017-10-28 04:37:02,435 [INFO] 2017-10-28 04:37:02,435 [INFO] info: report option not implemented, ignored.
info: report option not implemented, ignored.
2017-10-28 00:37:02,436 [INFO] sr_shovel t_dd2_f00 0001 stopping
running instance for config pelle_dd1_f04 (pid 15872) stopped.
running instance for config pelle_dd2_f05 (pid 15847) stopped.
    maximum of the shovels is: 1022

Then check show it went with flow_check.sh:

TYPE OF ERRORS IN LOG :

  1 /home/peter/.cache/sarra/log/sr_cpump_xvan_f14_001.log [ERROR] binding failed: server channel error 404h, message: NOT_FOUND - no exchange 'xcvan00' in vhost '/'
  1 /home/peter/.cache/sarra/log/sr_cpump_xvan_f15_001.log [ERROR] binding failed: server channel error 404h, message: NOT_FOUND - no exchange 'xcvan01' in vhost '/'


test  1 success: shovels t_dd1_f00 ( 1022 ) and t_dd2_f00 ( 1022 ) should have about the same number of items read
test  2 success: sarra tsarra (1022) should be reading about half as many items as (both) winnows (2240)
test  3 success: tsarra (1022) and sub t_f30 (1022) should have about the same number of items
test  4 success: max shovel (1022) and subscriber t_f30 (1022) should have about the same number of items
test  5 success: count of truncated headers (1022) and subscribed messages (1022) should have about the same number of items
test  6 success: count of downloads by subscribe t_f30 (1022) and messages received (1022) should be about the same
test  7 success: downloads by subscribe t_f30 (1022) and files posted by watch (1022) should be about the same
test  8 success: posted by watch(1022) and sent by sr_sender (1022) should be about the same
test  9 success: 1022 of 1022: files sent with identical content to those downloaded by subscribe
test 10 success: 1022 of 1022: poll test1_f62 and subscribe q_f71 run together. Should have equal results.
test 11 success: post test2_f61 1022 and subscribe r_ftp_f70 1021 run together. Should be about the same.
test 12 success: cpump both pelles (c shovel) should receive about the same number of messages (3665) (3662)
test 13 success: cdnld_f21 subscribe downloaded (1022) the same number of files that was published by both van_14 and van_15 (1022)
test 14 success: veille_f34 should post the same number of files (1022) that subscribe cdnld_f21 downloaded (1022)
test 15 success: veille_f34 should post the same number of files (1022) that subscribe cfile_f44 downloaded (1022)
test 16 success: Overall 15 of 15 passed!

blacklab%

If the flow_check.sh passes, then one has a reasonable confidence in the overall functionality of the python application, but the test coverage is not exhaustive. This is the lowest gate for committing changes to thy python code into the development branch. It is more qualitative sampling of the most common use cases rather than a thorough examination of all functionality. While not thorough, it is good to know the flows are working.

Note that the fclean subscriber looks at files in and keeps files around long enough for them to go through all the other tests. It does this by waiting a reasonable amount of time (45 seconds, the last time checked.) then it compares the file that have been posted by watch to the files created by downloading from it. As the sample now count proceeds, it prints “OK” if the files downloaded are identical to the ones posted by sr_watch. The addition of fclean and the corresponding cfclean for the cflow_test, are broken. The default setup which uses fclean and cfclean ensures that only a few minutes worth of disk space is used at a given time, and allows for much longer tests.

By default, the flow_test is only 1000 files, but one can ask it to run longer, like so:

./flow_limit.sh 50000

To accumulate fifty thousand files before ending the test. This allows testing of long term performance, especially memory usage over time, and the housekeeping functions of on_heartbeat processing.

Flow Cleanup

When done testing, run the ./flow_cleanup.sh script, which will kill the running servers and daemons, and delete all configuration files installed for the flow test, all queues, exchanges, and logs. This also needs to be done between each run of the flow test:

blacklab% ./flow_cleanup.sh
Stopping sr...
Cleanup sr...
Cleanup trivial http server...
web server stopped.
if other web servers with lost pid kill them
Cleanup trivial ftp server...
ftp server stopped.
if other ftp servers with lost pid kill them
Cleanup flow poster...
flow poster stopped.
if other flow_post.sh with lost pid kill them
Deleting queues:
Deleting exchanges...
Removing flow configs...
2018-02-10 14:17:34,150 [INFO] info: instances option not implemented, ignored.
2018-02-10 14:17:34,150 [INFO] info: report option not implemented, ignored.
2018-02-10 14:17:34,353 [INFO] info: instances option not implemented, ignored.
2018-02-10 14:17:34,353 [INFO] info: report option not implemented, ignored.
2018-02-10 09:17:34,837 [INFO] sr_poll f62 cleanup
2018-02-10 09:17:34,845 [INFO] deleting exchange xs_tsource_poll (tsource@localhost)
2018-02-10 09:17:35,115 [INFO] sr3_post shim_f63 cleanup
2018-02-10 09:17:35,122 [INFO] deleting exchange xs_tsource_shim (tsource@localhost)
2018-02-10 09:17:35,394 [INFO] sr3_post test2_f61 cleanup
2018-02-10 09:17:35,402 [INFO] deleting exchange xs_tsource_post (tsource@localhost)
2018-02-10 09:17:35,659 [INFO] sr_report tsarra_f20 cleanup
2018-02-10 09:17:35,659 [INFO] AMQP  broker(localhost) user(tfeed) vhost(/)
2018-02-10 09:17:35,661 [INFO] deleting queue q_tfeed.sr_report.tsarra_f20.89336558.04455188 (tfeed@localhost)
2018-02-10 09:17:35,920 [INFO] sr_report twinnow00_f10 cleanup
2018-02-10 09:17:35,920 [INFO] AMQP  broker(localhost) user(tfeed) vhost(/)
2018-02-10 09:17:35,922 [INFO] deleting queue q_tfeed.sr_report.twinnow00_f10.35552245.50856337 (tfeed@localhost)
2018-02-10 09:17:36,179 [INFO] sr_report twinnow01_f10 cleanup
2018-02-10 09:17:36,180 [INFO] AMQP  broker(localhost) user(tfeed) vhost(/)
2018-02-10 09:17:36,182 [INFO] deleting queue q_tfeed.sr_report.twinnow01_f10.48262886.11567358 (tfeed@localhost)
2018-02-10 09:17:36,445 [WARNING] option url deprecated please use post_base_url
2018-02-10 09:17:36,446 [WARNING] use post_base_dir instead of document_root
2018-02-10 09:17:36,446 [INFO] sr_sarra download_f20 cleanup
2018-02-10 09:17:36,446 [INFO] AMQP  broker(localhost) user(tfeed) vhost(/)
2018-02-10 09:17:36,448 [INFO] deleting queue q_tfeed.sr_sarra.download_f20 (tfeed@localhost)
2018-02-10 09:17:36,449 [INFO] exchange xpublic remains
2018-02-10 09:17:36,703 [INFO] sr_sender tsource2send_f50 cleanup
2018-02-10 09:17:36,703 [INFO] AMQP  broker(localhost) user(tsource) vhost(/)
2018-02-10 09:17:36,705 [INFO] deleting queue q_tsource.sr_sender.tsource2send_f50 (tsource@localhost)
2018-02-10 09:17:36,711 [INFO] deleting exchange xs_tsource_output (tsource@localhost)
2018-02-10 09:17:36,969 [INFO] sr_shovel t_dd1_f00 cleanup
2018-02-10 09:17:36,969 [INFO] AMQP  broker(dd.weather.gc.ca) user(anonymous) vhost(/)
2018-02-10 09:17:37,072 [INFO] deleting queue q_anonymous.sr_shovel.t_dd1_f00 (anonymous@dd.weather.gc.ca)
2018-02-10 09:17:37,095 [INFO] exchange xwinnow00 remains
2018-02-10 09:17:37,095 [INFO] exchange xwinnow01 remains
2018-02-10 09:17:37,389 [INFO] sr_shovel t_dd2_f00 cleanup
2018-02-10 09:17:37,389 [INFO] AMQP  broker(dd.weather.gc.ca) user(anonymous) vhost(/)
2018-02-10 09:17:37,498 [INFO] deleting queue q_anonymous.sr_shovel.t_dd2_f00 (anonymous@dd.weather.gc.ca)
2018-02-10 09:17:37,522 [INFO] exchange xwinnow00 remains
2018-02-10 09:17:37,523 [INFO] exchange xwinnow01 remains
2018-02-10 09:17:37,804 [INFO] sr_subscribe cclean_f91 cleanup
2018-02-10 09:17:37,804 [INFO] AMQP  broker(localhost) user(tsub) vhost(/)
2018-02-10 09:17:37,806 [INFO] deleting queue q_tsub.sr_subscribe.cclean_f91.39328538.44917465 (tsub@localhost)
2018-02-10 09:17:38,062 [INFO] sr_subscribe cdnld_f21 cleanup
2018-02-10 09:17:38,062 [INFO] AMQP  broker(localhost) user(tfeed) vhost(/)
2018-02-10 09:17:38,064 [INFO] deleting queue q_tfeed.sr_subscribe.cdnld_f21.11963392.61638098 (tfeed@localhost)
2018-02-10 09:17:38,324 [WARNING] use post_base_dir instead of document_root
2018-02-10 09:17:38,324 [INFO] sr_subscribe cfile_f44 cleanup
2018-02-10 09:17:38,324 [INFO] AMQP  broker(localhost) user(tfeed) vhost(/)
2018-02-10 09:17:38,326 [INFO] deleting queue q_tfeed.sr_subscribe.cfile_f44.56469334.87337271 (tfeed@localhost)
2018-02-10 09:17:38,583 [INFO] sr_subscribe clean_f90 cleanup
2018-02-10 09:17:38,583 [INFO] AMQP  broker(localhost) user(tsub) vhost(/)
2018-02-10 09:17:38,585 [INFO] deleting queue q_tsub.sr_subscribe.clean_f90.45979835.20516428 (tsub@localhost)
2018-02-10 09:17:38,854 [WARNING] extended option download_cp_command = ['cp --preserve=timestamps'] (unknown or not declared)
2018-02-10 09:17:38,855 [INFO] sr_subscribe cp_f61 cleanup
2018-02-10 09:17:38,855 [INFO] AMQP  broker(localhost) user(tsource) vhost(/)
2018-02-10 09:17:38,857 [INFO] deleting queue q_tsource.sr_subscribe.cp_f61.61218922.69758215 (tsource@localhost)
2018-02-10 09:17:39,121 [INFO] sr_subscribe ftp_f70 cleanup
2018-02-10 09:17:39,121 [INFO] AMQP  broker(localhost) user(tsource) vhost(/)
2018-02-10 09:17:39,123 [INFO] deleting queue q_tsource.sr_subscribe.ftp_f70.47997098.27633529 (tsource@localhost)
2018-02-10 09:17:39,386 [INFO] sr_subscribe q_f71 cleanup
2018-02-10 09:17:39,386 [INFO] AMQP  broker(localhost) user(tsource) vhost(/)
2018-02-10 09:17:39,389 [INFO] deleting queue q_tsource.sr_subscribe.q_f71.84316550.21567557 (tsource@localhost)
2018-02-10 09:17:39,658 [INFO] sr_subscribe t_f30 cleanup
2018-02-10 09:17:39,658 [INFO] AMQP  broker(localhost) user(tsub) vhost(/)
2018-02-10 09:17:39,660 [INFO] deleting queue q_tsub.sr_subscribe.t_f30.26453890.50752396 (tsub@localhost)
2018-02-10 09:17:39,924 [INFO] sr_subscribe u_sftp_f60 cleanup
2018-02-10 09:17:39,924 [INFO] AMQP  broker(localhost) user(tsource) vhost(/)
2018-02-10 09:17:39,927 [INFO] deleting queue q_tsource.sr_subscribe.u_sftp_f60.81353341.03950190 (tsource@localhost)
2018-02-10 09:17:40,196 [WARNING] option url deprecated please use post_base_url
2018-02-10 09:17:40,196 [WARNING] use post_broker to set broker
2018-02-10 09:17:40,197 [INFO] watch f40 cleanup
2018-02-10 09:17:40,207 [INFO] deleting exchange xs_tsource (tsource@localhost)
2018-02-10 09:17:40,471 [INFO] sr_winnow t00_f10 cleanup
2018-02-10 09:17:40,471 [INFO] AMQP  broker(localhost) user(tfeed) vhost(/)
2018-02-10 09:17:40,474 [INFO] deleting queue q_tfeed.sr_winnow.t00_f10 (tfeed@localhost)
2018-02-10 09:17:40,480 [INFO] deleting exchange xsarra (tfeed@localhost)
2018-02-10 09:17:40,741 [INFO] sr_winnow t01_f10 cleanup
2018-02-10 09:17:40,741 [INFO] AMQP  broker(localhost) user(tfeed) vhost(/)
2018-02-10 09:17:40,743 [INFO] deleting queue q_tfeed.sr_winnow.t01_f10 (tfeed@localhost)
2018-02-10 09:17:40,750 [INFO] deleting exchange xsarra (tfeed@localhost)
2018-02-10 14:17:40,753 [ERROR] config cno_trouble_f00 not found.
Removing flow config logs...
rm: cannot remove '/home/peter/.cache/sarra/log/sr_audit_f00.log': No such file or directory
Removing document root ( /home/peter/sarra_devdocroot )...
Done!

After the flow_cleanup.sh, to check that a test has completed, use:

sr3 status

which should show that there are no active configurations.

If the static_flow test works, then re-run the other tests: flakey_broker, transform_flow, and dynamic_flow.

Dynamic Flow Test Length

While most tests have a fixed duration, the dynamic flow test queries a remote server and can run for any length desired. The dynamic flow_test length defaults to 1000 files being flowed through the test cases. When in rapid development, one can supply an argument to shorten that:

./flow_limit.sh 200

Towards the end of a development cycle, longer flow_tests are adviseable:

./flow_limit.sh 20000

to identify more issues. sample run to 100,000 entries:

blacklab% ./flow_limit.sh 100000
initial sample building sample size 155 need at least 100000
sample now 100003 content_checks:GOOD missed_dispositions:0s:0
Sufficient!
stopping shovels and waiting...
2018-02-10 13:15:08,964 [INFO] 2018-02-10 13:15:08,964 [INFO] info: instances option not implemented, ignored.
info: instances option not implemented, ignored.
2018-02-10 13:15:08,964 [INFO] info: report option not implemented, ignored.
2018-02-10 13:15:08,964 [INFO] info: report option not implemented, ignored.
running instance for config pelle_dd2_f05 (pid 20031) stopped.
running instance for config pelle_dd1_f04 (pid 20043) stopped.
Traceback (most recent call last):ng...
  File "/usr/bin/rabbitmqadmin", line 1012, in <module>
    main()
  File "/usr/bin/rabbitmqadmin", line 413, in main
    method()
  File "/usr/bin/rabbitmqadmin", line 593, in invoke_list
    format_list(self.get(uri), cols, obj_info, self.options)
  File "/usr/bin/rabbitmqadmin", line 710, in format_list
    formatter_instance.display(json_list)
  File "/usr/bin/rabbitmqadmin", line 721, in display
    (columns, table) = self.list_to_table(json.loads(json_list), depth)
  File "/usr/bin/rabbitmqadmin", line 775, in list_to_table
    add('', 1, item, add_to_row)
  File "/usr/bin/rabbitmqadmin", line 742, in add
    add(column, depth + 1, subitem, fun)
  File "/usr/bin/rabbitmqadmin", line 742, in add
    add(column, depth + 1, subitem, fun)
  File "/usr/bin/rabbitmqadmin", line 754, in add
    fun(column, subitem)
  File "/usr/bin/rabbitmqadmin", line 761, in add_to_row
    row[column_ix[col]] = maybe_utf8(val)
  File "/usr/bin/rabbitmqadmin", line 431, in maybe_utf8
    return s.encode('utf-8')
AttributeError: 'float' object has no attribute 'encode'
maximum of the shovels is: 100008

While it is running one can run flow_check.sh at any time:

NB retries for sr_subscribe t_f30 0
NB retries for sr_sender 18

      1 /home/peter/.cache/sarra/log/sr_cpost_veille_f34_0001.log [ERROR] sr_cpost rename: /home/peter/sarra_devdocroot/cfr/observations/xml/AB/today/today_ab_20180210_e.xml cannot stat.
      1 /home/peter/.cache/sarra/log/sr_cpump_xvan_f14_0001.log [ERROR] binding failed: server channel error 404h, message: NOT_FOUND - no exchange 'xcvan00' in vhost '/'
      1 /home/peter/.cache/sarra/log/sr_cpump_xvan_f15_0001.log [ERROR] binding failed: server channel error 404h, message: NOT_FOUND - no exchange 'xcvan01' in vhost '/'
      1 /home/peter/.cache/sarra/log/sr_sarra_download_f20_0002.log [ERROR] Download failed http://dd2.weather.gc.ca//bulletins/alphanumeric/20180210/CA/CWAO/09/CACN00_CWAO_100857__WDK_10905
      1 /home/peter/.cache/sarra/log/sr_sarra_download_f20_0002.log [ERROR] Failed to reach server. Reason: [Errno 110] Connection timed out
      1 /home/peter/.cache/sarra/log/sr_sarra_download_f20_0002.log [ERROR] Download failed http://dd2.weather.gc.ca//bulletins/alphanumeric/20180210/CA/CWAO/09/CACN00_CWAO_100857__WDK_10905. Type: <class 'urllib.error.URLError'>, Value: <urlopen error [Errno 110] Connection timed out>
      1 /home/peter/.cache/sarra/log/sr_sarra_download_f20_0004.log [ERROR] Download failed http://dd2.weather.gc.ca//bulletins/alphanumeric/20180210/SA/CYMM/09/SACN61_CYMM_100900___53321
      1 /home/peter/.cache/sarra/log/sr_sarra_download_f20_0004.log [ERROR] Failed to reach server. Reason: [Errno 110] Connection timed out
      1 /home/peter/.cache/sarra/log/sr_sarra_download_f20_0004.log [ERROR] Download failed http://dd2.weather.gc.ca//bulletins/alphanumeric/20180210/SA/CYMM/09/SACN61_CYMM_100900___53321. Type: <class 'urllib.error.URLError'>, Value: <urlopen error [Errno 110] Connection timed out>
      1 /home/peter/.cache/sarra/log/sr_sarra_download_f20_0004.log [ERROR] Download failed http://dd2.weather.gc.ca//bulletins/alphanumeric/20180210/CS/CWEG/12/CSCN03_CWEG_101200___12074
more than 10 TYPES OF ERRORS found... for the rest, have a look at /home/peter/src/sarracenia/test/flow_check_errors_logged.txt for details

test  1 success: shovels t_dd1_f00 (100008) and t_dd2_f00 (100008) should have about the same number of items read
test  2 success: sarra tsarra (100008) should be reading about half as many items as (both) winnows (200016)
test  3 success: tsarra (100008) and sub t_f30 (99953) should have about the same number of items
test  4 success: max shovel (100008) and subscriber t_f30 (99953) should have about the same number of items
test  5 success: count of truncated headers (100008) and subscribed messages (100008) should have about the same number of items
test  6 success: count of downloads by subscribe t_f30 (99953) and messages received (100008) should be about the same
test  7 success: same downloads by subscribe t_f30 (199906) and files posted (add+remove) by watch (199620) should be about the same
test  8 success: posted by watch(199620) and subscribed cp_f60 (99966) should be about half as many
test  9 success: posted by watch(199620) and sent by sr_sender (199549) should be about the same
test 10 success: 0 messages received that we don't know what happenned.
test 11 success: sarra tsarra (100008) and good audit 99754 should be the same.
test 12 success: poll test1_f62 94865 and subscribe q_f71 99935 run together. Should have equal results.
test 13 success: post test2_f61 99731 and subscribe r_ftp_f70 99939 run together. Should be about the same.
test 14 success: posts test2_f61 99731 and shim_f63 110795 Should be the same.
test 15 success: cpump both pelles (c shovel) should receive about the same number of messages (160737) (160735)
test 16 success: cdnld_f21 subscribe downloaded (50113) the same number of files that was published by both van_14 and van_15 (50221)
test 17 success: veille_f34 should post twice as many files (100205) as subscribe cdnld_f21 downloaded (50113)
test 18 success: veille_f34 should post twice as many files (100205) as subscribe cfile_f44 downloaded (49985)
test 19 success: Overall 18 of 18 passed (sample size: 100008) !

blacklab%

This test was fired up at the end of the day, as it takes several hours, and results examined the next morning.

High volume sample

Trying the flow test with higher volume of notification messages (ie. 100 000) is one step closer to the goal of having a flow test running continously. This is motivated by our testing purposes.

Limitation

Ubuntu have a limitation that tops inotify watches and that we encountered in #204 . We can overcome this by setting the related sysctl variable. First, check what is the limit of your system:

$ sysctl fs.inotify.max_user_watches
fs.inotify.max_user_watches = 8196

If the limit is too low (ie. 8196), change it to a more appropriate level for the flow test:

$ sudo sysctl fs.inotify.max_user_watches=524288

To make this change permanent add this line to /etc/sysctl.conf:

fs.inotify.max_user_watches=524288

Then excute sysctl -p and the system should now support high volume of inotify events.

Flow Test Stuck

Sometimes flow tests (especially for large numbers) get stuck because of problems with the data stream (where multiple files get the same name) and so earlier versions remove later versions and then retries will always fail. Eventually, we will succeed in cleaning up the dd.weather.gc.ca stream, but for now sometimes a flow_check gets stuck ‘Retrying.’ The test has run all the notification messages required, and is at a phase of emptying out retries, but just keeps retrying forever with a variable number of items that never drops to zero.

To recover from this state without discarding the results of a long test, do:

^C to interrupt the flow_check.sh 100000
blacklab% sr3 stop
blacklab% cd ~/.cache/sarra
blacklab% ls */*/*retry*
shovel/pclean_f90/sr_shovel_pclean_f90_0001.retry        shovel/pclean_f92/sr_shovel_pclean_f92_0001.retry        subscribe/t_f30/sr_subscribe_t_f30_0002.retry.new
shovel/pclean_f91/sr_shovel_pclean_f91_0001.retry        shovel/pclean_f92/sr_shovel_pclean_f92_0001.retry.state
shovel/pclean_f91/sr_shovel_pclean_f91_0001.retry.state  subscribe/q_f71/sr_subscribe_q_f71_0004.retry.new
blacklab% rm */*/*retry*
blacklab% sr3 start
blacklab%
blacklab%  ./flow_check.sh 100000
Sufficient!
stopping shovels and waiting...
2018-04-07 10:50:16,167 [INFO] sr_shovel t_dd2_f00 0001 stopped
2018-04-07 10:50:16,177 [INFO] sr_shovel t_dd1_f00 0001 stopped
2018-04-07 14:50:16,235 [INFO] info: instances option not implemented, ignored.
2018-04-07 14:50:16,235 [INFO] info: report option not
implemented, ignored.
2018-04-07 14:50:16,235 [INFO] info: instances option not implemented, ignored.
2018-04-07 14:50:16,235 [INFO] info: report option not
implemented, ignored.
running instance for config pelle_dd1_f04 (pid 12435) stopped.
running instance for config pelle_dd2_f05 (pid 12428) stopped.
maximum of the shovels is: 100075


blacklab% ./flow_check.sh

                 | dd.weather routing |
test  1 success: sr_shovel (100075) t_dd1 should have the same number
of items as t_dd2 (100068)
test  2 success: sr_winnow (200143) should have the sum of the number
of items of shovels (200143)
test  3 success: sr_sarra (98075) should have the same number of items
as winnows'post (100077)
test  4 success: sr_subscribe (98068) should have the same number of
items as sarra (98075)
                 | watch      routing |
test  5 success: watch (397354) should be 4 times subscribe t_f30 (98068)
test  6 success: sr_sender (392737) should have about the same number
of items as watch (397354)
test  7 success: sr_subscribe u_sftp_f60 (361172) should have the same
number of items as sr_sender (392737)
test  8 success: sr_subscribe cp_f61 (361172) should have the same
number of items as sr_sender (392737)
                 | poll       routing |
test  9 success: sr_poll test1_f62 (195408) should have half the same
number of items of sr_sender(196368)
test 10 success: sr_subscribe q_f71 (195406) should have about the
same number of items as sr_poll test1_f62(195408)
                 | flow_post  routing |
test 11 success: sr3_post test2_f61 (193541) should have half the same
number of items of sr_sender(196368)
test 12 success: sr_subscribe ftp_f70 (193541) should have about the
same number of items as sr3_post test2_f61(193541)
test 13 success: sr3_post test2_f61 (193541) should have about the same
number of items as shim_f63 195055
                 | py infos   routing |
test 14 success: sr_shovel pclean_f90 (97019) should have the same
number of watched items winnows'post (100077)
test 15 success: sr_shovel pclean_f92 (94537}) should have the same
number of removed items winnows'post (100077)
test 16 success: 0 messages received that we don't know what happenned.
test 17 success: count of truncated headers (98075) and subscribed
messages (98075) should have about the same number of items
                 | C          routing |
test 18 success: cpump both pelles (c shovel) should receive about the
same number of messages (161365) (161365)
test 19 success: cdnld_f21 subscribe downloaded (47950) the same
number of files that was published by both van_14 and van_15 (47950)
test 20 success: veille_f34 should post twice as many files (95846) as
subscribe cdnld_f21 downloaded (47950)
test 21 success: veille_f34 should post twice as many files (95846) as
subscribe cfile_f44 downloaded (47896)
test 22 success: Overall 21 of 21 passed (sample size: 100077) !

NB retries for sr_subscribe t_f30 0
NB retries for sr_sender 36

So, in this case, the results are still good in spite of not quite being able to terminate. If there was a significant problem, the cumulation would indicate it.

Flow tests with MQTT

Flow tests can be run where certain components use the MQTT protocol, instead of AMQP.

FIXME: steps missing, more clarity required.

MQTT broker is installed
the bunnymaster tsource, tfeed, tsub users defined and given passwords (broker dependent.)
for each user: an mqtt://user:pw@brokerhost url’s line is added to ~/.config/sr3/credentials.conf
edit the variable MQP in ~/.config/sr3/default.conf, MQP is used by the flow tests.

Most components will use MQTT instead of amqp and can be run normally.

Commits to the Development Branch

Aside from typos, language fixups in the documentation, and incrementing the version, developers are not expected to commit to stable. All work happens on development branches, and all testing is expected to pass before one considers affecting stable. Once the branch development is complete, or a unit of work-in-progress is felt to be worth merging to stable, one must summarize the changes from the branch for the debian change log, request on github.

git checkout issueXXX  # v02_issueXXX for v2 work., github suggested branch names are fine also.
vi CHANGES.rst # summarize the changes in Restructured Text
dch # copy/paste from CHANGES.rst, inserting one leading space.
vi doc/UPGRADING.rst # rarely, if code has user impact.
vi doc/fr/UPGRADING.rst # bon... ceci est visible aux usagers, donc...
git commit -a
git push
# issue a pull request on github.com.

A Second developer will review the pull request and the reviewer will decide on whether merging is appropriate. The developer is expected to examine each commit, and understand it to some degree.

If the pull-request has one of the following (substantial changes, new functionality, modifications to critical code structure) , it is recommended to have a Third developer also review the pull request. The expectation from this developer are the same as from the previous.

The github Actions looks at pull requests and will flow tests on them. If the tests pass, then that is good qualitative indicator, however the tests are a bit fragile at the moment, so if they fail, it would be ideal for the reviewer to run the tests in their own development environment. If it passes in the local developer environment one can approve a merge in spite of Github Actions’ complaints.

Key Branches

There is a long running discussion about Which Version is stable The current set up is that there are four principal branches:

stable branch is the release version of sr3, merging from pre-release. used to build sr3 packages in the MetPX repository.
development … The version 3 work in progress branch is a next version of sarracenia. the development branch is used to build sr3 packages for the Daily and Pre-Release repositories on launchpad.net.
stable_py36 and pre-relrease-36 are branched from stable and pre_release respectively to adjust for building packages on older operating systems that have older versions of python (and no support for hatchling.)
issue branches to be merged to development, it should be start with issueXXX or suggested branch names from github are ok also.
sometimes, multiple branches are needed for a single issue, say for variations of a fix, eg. issueXXX_2_do_it_this_way .
v2_dev … the integration branch for v2 maintenance used prior to promotion to v2_stable.
v2_stable … generally this branch gets code via merges from v2_dev, after the pre-release has been tested on a as many systems as possible. used to build packages on the stable: MetPX

Repositories

For Ubuntu operating systems, the launchpad.net site is the best way to provide packages that are fully integrated ( built against current patch levels of all dependencies (software components that Sarracenia relies on to provide full functionality.)) Ideally, when running a server, a one should use one of the repositories, and allow automated patching to upgrade them as needed.

Repositories:

Daily https://launchpad.net/~ssc-hpc-chp-spc/+archive/ubuntu/metpx-daily (living on the edge… ) automated daily build of sr3 packages happens from development branch.
Pre-Release https://launchpad.net/~ssc-hpc-chp-spc/+archive/ubuntu/metpx-pre-release (for newest features.) from development branch. Developers manually trigger builds here when it seems appropriate (testing out code that is ready for release.)
Release https://launchpad.net/~ssc-hpc-chp-spc/+archive/ubuntu/metpx (for maximum stability) from v2_stable branch. After testing in systems subscribed to pre-releases, Developers merge from v2_dev branch into v2_stable one, and manually trigger a build.

for more discussion see Which Version is stable

Building a Release

MetPX-Sarracenia is distributed in a few different ways, and each has it’s own build process. Packaged releases are always preferable to one off builds, because they are reproducible.

When development requires testing across a wide range of servers, it is preferred to make an alpha release, rather than installing one off packages. So the preferred mechanisms is to build the ubuntu and pip packages at least, and install on the test machines using the relevant public repositories.

To publish a release one needs to:

Set the version.
upload the release to pypi.org so that installation with pip succeeds.
upload the release to launchpad.net, so that the installation of debian packages using the repository succeeds.
upload the release notes and binaries to github.com.
increment the version for future commits to v2_dev or development.

Versioning Scheme

Each release will be versioned as <protocol version>.<YY>.<MM> <segment>

Where:

protocol version is the message version. In Sarra notification messages, they are all prefixed with v02 (at the moment).
YY is the last two digits of the year of the initial release in the series.
MM is a TWO digit month number i.e. for April: 04.
segment is what would be used within a series. from pep0440: X.YaN # Alpha release X.YbN # Beta release X.YrcN # Release Candidate X.Y # Final release X.ypostN #ack! patched release.

Currently, 3.00 is still stabilizing, so the year/month convention is not being applied. Releases are currently 3.00.iircj where:

ii – incremental number of pre-releases of 3.00

j – beta increment.

At some point 3.00 will be complete & solid enough that the we will resume the year/month convention, hopefully 3.22.

Releases are classified as follows:

Alpha: Snapshot releases taken directly from development, with no other qualitative guarantees. No guarantee of functionality, some components may be partially implemented, some breakage may occur. No bug-fixes, issues addressed by subsequent version. Often used for early end-to-end testing (rather than installing custom from tree on each test machine.)
Beta: Feature Complete for a given release. Components in their final form for this release. Documentation exists in at least one language. All previously known release block bugs addressed. No bug-fixes, issues addressed by subsequent version.
RC - Release Candidate.: Implies it has gone through beta to identify and address major issues. Translated documentation available. No bug-fixes, issues addressed by subsequent version.

Final versions have no suffix and are considered stable and supported. Stable should receive bug-fixes if necessary from time to time. One can build python wheels, or debian packages for local testing purposes during development.

Note

If you change default settings for exchanges / queues as part of a new version, keep in mind that all components have to use the same settings or the bind will fail, and they will not be able to connect. If a new version declares different queue or exchange settings, then the simplest means of upgrading (preserving data) is to drain the queues prior to upgrading, for example by setting, the access to the resource will not be granted by the server. (??? perhaps there is a way to get access to a resource as is… no declare) (??? should be investigated)

Changing the default requires the removal and recreation of the resource. This has a major impact on processes…

Example:

The first alpha release in January 2016 would be versioned as metpx-sarracenia-2.16.01a01

Set the Version

This is done to start development on a version. It should be done on development after every release.

git checkout development
Edit sarracenia/_version.py manually and set the version number.
Edit CHANGES.rst to add a section for the version.
run dch to start the changelog for the current version. * change unstable to UNRELEASED (maybe done automatically by dch.)
git commit -a
git push

If development continues and the time passes without the release occurring, then the version needs to be set to again (or overwritten). For example, a development cycle begins in August, the version in stable will be 3.19.08b1… but if development continues into September, one should use this procedure the change the version to 3.19.09b1.

note on Debian terminology:

unstable … new software brought into the OS stablization tree.
testing … software considered stable enough to be included in the next “stable” release.
stable … part of a stable Debian release: frozen software that will not be patched except for security issues.

for Metpx-Sr3, we have not yet approached the longevity of versions implied by Debian stable. The labels don’t refer to how stable the software itself is, merely to how often the software changes (increments versions.) More info: Debian Version History so when performing releases, the changelog is changed from UNRELEASED to unstable, and back to UNRELEASED when working between releases, as per Debian custom.

Release Versions and Quality Assurance

The Quality Assurance process, or QA occurs mainly on the development branch. prior to accepting a release, and barring known exceptions,

QA tests automatically keyed to pushes to the development branch should all pass. (All related github actions.) tests: static, no_mirror, flakey_broker, restart_server, dynamic_flow are included in “flow.yml”
build an ubuntu 18.04 vm and run the flow tests there to ensure that it works. (installation method: cloning from development on github.) tests: static, no_mirror, flakey_broker, restart_server, dynamic_flow
build a redhat 8 vm and run the flow test there to ensure that it works. (installation method: cloning from development on github.) tests: static, no_mirror, flakey_broker, restart_server, dynamic_flow
After that, a pre-release version is prepared. It should have an rcX suffix publish to: https://launchpad.net/~ssc-hpc-chp-spc/+archive/ubuntu/metpx-pre-release publish to: pypi.org (rcX suffix is detected as a pre-release.) publish to: github.com (tag as pre-release.)

the pre-release repository.
- pre-release version is installed on development server. (from launchpad repo.)
- pre-release version is installed on some staging workloads. i(from launchpad repo.)
- if the pre-release version runs all configured flows without incidents for a week.
Then a stable version is declared, pushed to the stable repository:
- ensure there is no rcX suffix on the version.
- publish to: https://launchpad.net/~ssc-hpc-chp-spc/+archive/ubuntu/metpx
- publish to: pypi.org (remove rcX suffix so it is considered latest.)
- publish to: github.com (tag as latest .)

the mechanics of preparing each release are documented in the following section.

For extensive discussion see: https://github.com/MetPX/sarracenia/issues/139

Releasing

Prior to releasing, ensure that all QA tests in the section above are passed. When development for a version is complete. The following should occur:

A tag should be created to identify the end of the cycle:

git checkout development
git tag -a sarra-v3.16.01a01 -m "release 3.16.01a01"
git push
git push origin sarra-v3.16.01a01

Once the tag is in the development branch, promote it to stable:

git checkout stable
git merge development
git push

Once stable is updated on github, the docker images will be automatically upgraded, but we then need to update the various distribution methods: PyPI, and Launchpad

Once package generation is complete, one should Set the Version in development to the next logical increment to ensure no further development occurs that is identified as the released version.

PyPi

Pypi does not distinguish between older and newer python releases. There is only one package version for all supported versions. When uploading from a new OS, the versions in use on the OS are inferred to be the minimum, and so installation on older operating systems may be blocked by generated dependencies on overly modern versions.

So when uploading to pypi, always do so from the oldest operating system where it needs to work. upward compatibility is more likely than downward.

Pypi Credentials go in ~/.pypirc. Sample Content:

[pypi]
username: SupercomputingGCCA
password: <get this from someone>

Assuming pypi upload credentials are in place, uploading a new release used to be a one liner:

python3 setup.py bdist_wheel upload

on older systems, or on (python >= 3.7) newer ones:

python3 -m build --no-isolation
twine upload dist/metpx_sarracenia-2.22.6-py3-none-any.whl dist/metpx_sarracenia-2.22.6.tar.gz

Should always include source (the .tar.gz file) Note that the CHANGES.rst file is in restructured text and is parsed by pypi.python.org on upload.

Note

When uploading pre-release packages (alpha,beta, or RC) PYpi does not serve those to users by default. For seamless upgrade, early testers need to do supply the --pre switch to pip:

pip3 install --upgrade --pre metpx-sarracenia

On occasion you may wish to install a specific version:

pip3 install --upgrade metpx-sarracenia==2.16.03a9

command line use of setup.py is deprecated. Replaced by build and twine.

Launchpad

Canonical, the company behind Ubuntu, provides Launchpad as a means of enabling third parties to build packages for their operating system releases. It turns out that the newer OS versions have dependencies that are not available on the old ones. So the development branch is configured to build on newer releases, but an a separate branch must be created when creating releases for ubuntu bionic (18.04) and focal (20.04)

git checkout stable
git branch -D stable_focal_bionic
git branch stable_focal_bionic
git checkout stable_focal_bionic
vi debian/control
# remove pybuild-plugin-pyproject from the "Build-Depends"
git rm pyproject.toml
# remove the new-style installer to force use of setup.py
git commit -a -m "adjust for older os"

There might be a “–force” required at some point. Perhaps something along the lines of:

git push origin stable_focal_bionic --force

Then proceed with Launchpad instructions.

Automated Build

Ensure the code mirror is updated by checking the Import details by checking this page for sarracenia
if the code is out of date, do Import Now , and wait a few minutes while it is updated.
once the repository is upto date, proceed with the build request.
Go to the sarracenia release recipe
Go to the sr3 release recipe
Click on the Request build(s) button to create a new release
for Sarrac, follow the procedure here
The built packages will be available in the metpx ppa

Building a Windows Installer

One can also build a Windows installer with that script. It needs to be run from a Linux OS (preferably Ubuntu 18) in the root directory of Sarracenia’s git. find the python version in use:

fractal% python -V
Python 3.10.12
fractal%

So this is python 3.10. Only a single minor version will have the embedded package needed by pynsist to build the executable, so look at:

https://www.python.org/downloads/windows/

Then go look on python.org, for the “right” version (for 3.10, it is 3.10.11 ) Then, from the shell, run:

sudo apt install nsis
pip3 install pynsist wheel
./generate-win-installer.sh 3.10.11 2>&1 > log.txt

The final package will be generated into build/nsis directory.

Daily Builds

Daily builds are configured using this recipe for python and this recipe for C and are run once per day when changes to the repository occur. These packages are stored in the metpx-daily ppa. One can also Request build(s) on demand if desired.

Manual Process

The process for manually publishing packages to Launchpad ( https://launchpad.net/~ssc-hpc-chp-spc ) involves a more complex set of steps, and so the convenient script publish-to-launchpad.sh will be the easiest way to do that. Currently the only supported releases are trusty and xenial. So the command used is:

publish-to-launchpad.sh sarra-v2.15.12a1 trusty xenial

However, the steps below are a summary of what the script does:

for each distribution (precise, trusty, etc) update debian/changelog to reflect the distribution
build the source package using:
```
debuild -S -uc -us
```
sign the .changes and .dsc files:
```
debsign -k<key id> <.changes file>
```

upload to launchpad:

dput ppa:ssc-hpc-chp-spc/metpx-<dist> <.changes file>

Note: The GPG keys associated with the launchpad account must be configured in order to do the last two steps.

Backporting a Dependency

Example:

backportpackage -k<key id> -s bionic -d xenial -u ppa:ssc-hpc-chp-spc/ubuntu/metpx-daily librabbitmq

Building an RPM

This process is currently a bit clumsy, but it can provide usable RPM packages. Example of creating a multipass image for fedora to build with:

fractal% multipass launch -m 8g --name fed34 https://mirror.csclub.uwaterloo.ca/fedora/linux/releases/34/Cloud/x86_64/images/Fedora-Cloud-Base-34-1.2.x86_64.raw.xz
Launched: fed34
fractal%

Based on https://www.redhat.com/sysadmin/create-rpm-package … install build-dependencies:

sudo dnf install -y rpmdevtools rpmlint git
git clone -b development https://github.com/MetPX/sarracenia sr3
cd sr3

One can build a very limited sort of rpm package on an rpm based distro by The names of the package for file magic data (to determin file types) has different names on ubuntu vs. redhat. The last three lines of dependencies in pyproject.toml are about <<<<<<< HEAD “python-magic”, but on redhat/fedora >= 9, it needs to be “file-magic” instead:

# remove last three lines of dependencies in setup.py

* on redhat <=8: vi setup.py ;  python3 setup.py bdist_rpm

# might work, but might need some removals also.
* on redhat >=9: vi pyproject.toml;  python3 -m build

“python-magic”, but on redhat, it needs to be “file-magic” instead:

vi pyproject.toml

using the normal (for Redhat) rpmbuild tool:

rpmbuild --build-in-place -bb metpx-sr3.spec

When doing this on the redhat 8, edit the metpx-sr3.spec and potentially pyproject.toml to remove the other dependencies because there are no OS packages for: paramiko, watchdog, xattr, & magic. Eventually, one will have removed enough that the rpm file will be built.

One can check if the dependencies are there like so:

[ubuntu@fed39 sr3]$ rpm -qR /home/ubuntu/rpmbuild/RPMS/noarch/metpx-sr3-3.00.47-0.fc39.noarch.rpm

/usr/bin/python3
python(abi) = 3.12
python3-appdirs
python3-humanfriendly
python3-humanize
python3-jsonpickle
python3-paramiko
python3-psutil
python3-xattr
python3.12dist(appdirs)
python3.12dist(humanfriendly)
python3.12dist(humanize)
python3.12dist(jsonpickle)
python3.12dist(paramiko)
python3.12dist(psutil) >= 5.3
python3.12dist(watchdog)
python3.12dist(xattr)
rpmlib(CompressedFileNames) <= 3.0.4-1
rpmlib(FileDigests) <= 4.6.0-1
rpmlib(PartialHardlinkSets) <= 4.0.4-1
rpmlib(PayloadFilesHavePrefix) <= 4.0-1
rpmlib(PayloadIsZstd) <= 5.4.18-1

[ubuntu@fed39 sr3]$

You can see all of the prefixed python3 dependencies required, as well as the recommended binary accellerator packages are listed. Then if you install with dnf install, it will pull them all in. Unfortunately, this method does not allow the specification of version of the python dependencies which are stripped out. on Fedora 34, this is not a problem, as all versions are new enough. Such a package should install well.

After installation, one can supplement, installing missing dependencies using pip (or pip3.) Can check how much sr3 is working using sr3 features and use pip to add more features after the RPM is installed.

Certainly a bit clumsy, and not sure if it will work with older versions: Help Wanted

github

Click on Releases, Edit the release:

should we have release names?
copy/paste of changes from CHANGES.rst
copy/paste of Installation bit at the end from a previous release.
build packages locally or download from other sources. drag and drop into the release.

This will give us the ability to have old versions available. launchpad.net doesn’t seem to keep old versions around.

Development Environment

Local Python

Working with a non-packaged version:

notes:

pip install -e .

Windows

Install winpython from github.io version 3.5 or higher. Then use pip to install from PyPI.

Conventions

Below are some coding practices that are meant to guide developers when contributing to sarracenia. They are not hard and fast rules, just guidance.

When to Report

sr_report(7) notification messages should be emitted to indicate final disposition of the data itself, not any notifications or report messages (don’t report report messages, it becomes an infinite loop!) For debugging and other information, the local log file is used. For example, sr_shovel does not emit any sr_report(7) messages, because no data is transferred, only messages.

Adding a New Dependency

Dependency Management is a complicated topic, because python has many different installation methods into disparate environments, and Sarracenia is multi-platform. Standard python practice for dependencies is to make them required by listing them in requirements.txt or setup.py, and require all users to install them. In most python applications, if a dependency is missing, it just crashes with a import failure message of some kind.

In Sr3, we have found that there are many different environments being deployed into where satisfying dependencies can be more trouble than they are worth, so each of the dependencies in setup.py are also dealt with in sarracenia/featuredetection, and the feature detection code allows the application to keep working, just without the functionality provided by the missing module. This is called degradation or degraded mode. The idea being to help the user do as much as they can, in the environment they have, while telling them what is missing, and what would ideally be added.

for a full discussion see:

Managing Dependencies (Discussion) <https://github.com/MetPX/sarracenia/issues/741>

Short version:

In addition to requirements.dev/setup.py, if you need to add a new library that isn’t part of batteries included, typically provided by a separate os or pip package, then you want to provide for the package to still work in the event that the package is not available (albeit without the function you are adding) and to add support for explaining what’s missing using the sarracenia/featuredetection.py module.

In that module is a features data structure, where you add an entry explaining the import needed, and the functionality it brings to Sr3. You also add if feature[‘x’][‘present’] guards in the code where you are using the feature, in order to allow the code to degrade elegantly.

If the dependency is added in a plugin, then there is also a method for that described here:

Plugin Developer Guide <../Explanation/SarraPluginDev.html#callbacks-that-need-python-modules>