you've also checked that, as the rabbitmq user, it can access all of those files?
Aha, restarting gives
Interface: 127.0.0.1, port: 5672, protocol: amqp, purpose: AMQP 0-9-1 and AMQP 1.0
Interface: 0.0.0.0, port: 5671, protocol: amqp/ssl, purpose: AMQP 0-9-1 and AMQP 1.0 over TLS
so worth the dumb question 😅 I assumed rerunning the playbook would automatically restart it
Ahhh no, it won't unless it detects a configuration change (which didn't happen)
nonetheless
$ curl https://localhost:5671 -k
curl: (35) OpenSSL SSL_connect: Connection reset by peer in connection to localhost:5671
1 reply
https://galaxyproject.github.io/gxadmin/#/README.galaxy?id=galaxy-amqp-test might also be useful for you, at some point.
is there any chance it's working despite that error? does galaxy/pulsar talk to it now?
I guess the relevant bit is
Process <0.993.0> with 0 neighbours crashed with reason: no function clause matching tls_handshake_1_3:
I am getting an error that my tool - lukasa (https://github.com/pvanheus/lukasa/tree/main/galaxy) doesn't exist in the testtoolshed . On the other hand, if I try and create an entry for it I'm told it exists. Any ideas what might be going on?
1 reply
and if there is anything I can do to prod the system? I guess this is more of a tool authoring question but also wonder if it is a bug in the toolshed itself
natefoo (Nate Coraor) can you please check the sanbi-uwc lukasa repo on testtoolshed to see if there is something wrong with it. thanks
trying to update it
fixed
You can set
virtualenv in the Gravity config which should ensure that it is explicitly prepended to the commands
Hello, I am having issues with
welcome_url. I have a custom HTML placed in galaxy_server_dir/static/welcome.html and I configured the same in galaxy.yml. I am serving galaxy with a prefix cos and using Nginx as reverse proxy. Currently I am testing the setup locally inside a podman container with self signed certificates. When I access the welcome page at https://localhost:8443/cos, nginx is redirecting https://localhost:8443/cos/welcome to http://localhost:8443/cos/static/welcome.html. Notice the change of scheme from HTTPS to HTTP. Does anyone have any idea why is it changing the scheme? I am attaching a screenshot with browser network data. Cheers
1 reply
Hi all -- would someone be able to help with this? https://help.galaxyproject.org/t/galaxy-gunicorn-cant-connect-to-socket-file/8851 Thanks!
Thanks Jennifer Hillman-Jackson , I definitely will need help with "galaxy gunicorn can't connect to socket file". The socket is up and listening, but gunicorn is unable to connect: $ sudo systemctl status gunicorn-galaxy.socket
● gunicorn-galaxy.socket - gunicorn socket for galaxy
Loaded: loaded (/etc/systemd/system/gunicorn-galaxy.socket; enabled; vendor preset: enabled)
Active: active (listening) since Mon 2022-10-17 20:55:47 UTC; 51min ago
Triggers: ● gunicorn-galaxy.service
Listen: /run/gunicorn-galaxy.sock (Stream)
CGroup: /system.slice/gunicorn-galaxy.socket
Oct 17 20:55:47 ip-10-0-1-180 systemd[1]: Listening on gunicorn socket for galaxy.
These are the few things I've tried to get rid of "Can't connect to /run/gunicorn-galaxy.sock" file error:
change ownership of gunicorn-galaxy.sock to galaxy:galaxy => FAIL
- single quotes on gunicorn bind def in galaxy.yml => FAIL
- single quotes on gunicorn bind def in galaxy/database/gravity/galaxy.yml => FAIL
- update Systemd config to match exactly https://docs.galaxyproject.org/en/latest/admin/scaling.html => FAIL
- add flags to Systemd config as described in https://unix.stackexchange.com/questions/573980/cant-open-socket-im-systemd-service => FAIL
- restart gunicorn-galaxy.socket => FAIL
I also tried setting systemd as described in : https://training.galaxyproject.org/training-material/topics/admin/tutorials/stop-worrying-love-systemd/tutorial.html => FAIL: INFO spawnerr: can't find command 'celery'
I then reverted to the systemd config described in the Galaxy Project 22.09.dev0 docs, but adding UMask=022 => FAIL Can't connect to /run/gunicorn-galaxy.sock
I then reverted to the systemd config described in the Galaxy Project 22.09.dev0 docs, but adding UMask=022 => FAIL Can't connect to /run/gunicorn-galaxy.sock
FYI I've had no problem setting up Django apps with gunicorn + systemd + NGINX. I'm wondering if the problem with galaxy is that there is no explicit '--bind <path-to-my-socket-file>' option that appears in the ExecStart line of the dot service file?
26 replies
The are Galaxy servers in all timezones, but people may just be busy
Hi Folks,
I'm having some challenges with Python Dynamic Destinations and the various Galaxy functions/classes...
Background: For RNASTAR, I want to request sufficient memory from our compute nodes based on the reference genome size (e.g. Human hg38 = ~3 Gbp, Memory = ~32 Gb). At this point in time, it calculates a "nativeSpecification" string with the resources I need.
Problem: I am having difficulty with my function and returning the job destination. I want to take an existing JobDestination (retrieved by calling
Question: How can I retrieve the existing "rnastar" destination settings, make the changes I need (e.g. changing
I KNOW I am missing something small, but I cannot find it.
Thanks folks,
Darren Cullerne
I'm having some challenges with Python Dynamic Destinations and the various Galaxy functions/classes...
Background: For RNASTAR, I want to request sufficient memory from our compute nodes based on the reference genome size (e.g. Human hg38 = ~3 Gbp, Memory = ~32 Gb). At this point in time, it calculates a "nativeSpecification" string with the resources I need.
Problem: I am having difficulty with my function and returning the job destination. I want to take an existing JobDestination (retrieved by calling
app.job_config.get_destination("rnastar")), modify it locally, then execute my job with the modified JobDestination. So far, I've worked out that return 'rnastar' sends it to an existing job destination, and return JobDestination(id="rnastar_mod", runner="local") sends it to a new job destination, but none of the parameters I need are there (unless I build the whole thing from scratch).Question: How can I retrieve the existing "rnastar" destination settings, make the changes I need (e.g. changing
id or params) then execute my job using this modified JobDestination?I KNOW I am missing something small, but I cannot find it.
Thanks folks,
Darren Cullerne
3 replies
Hello, I have a question about the usage of
external_chown_script.py in cluster jobs. So, galaxy user (the user that is running Galaxy process) will give the ownership of the directory created for the job to the real user, run the job and once the job finishes, galaxy user will take back the ownership of the directory. Is it done to give the real user permissions to create files in the directory? If I have all the real users that use Galaxy in one group, say galaxyusers, and if I set the group ownership of jobs directory to this galaxyusers group with rw permissions, the real users will be able to write to this directory without any chown operation. At the same time galaxy user would be able to delete the folder once it finishes the job. Would this approach work? What I am not sure is does Galaxy moves any files to elsewhere once the job finishes before deleting the job directory? Thanks!!
I'm exploring the cli and python lessons in GTN and none of the notebook downloads I've tried want to open in JupyterLab running interactively on usegalaxy.eu - complain of JSON format error, anyone else had this problem?
Mh, strange can you share a history when this happens? Starting a Jupyter instance wo notebook works?
The simplest repro I can find is trying to launch one of the notebooks in jupyterlite: https://training.galaxyproject.org/training-material/jupyter/lab/index.html?path=data-science-python-basics.ipynb
I can start a jupyter instance on usegalaxy.eu without issue
problem comes once I have downloaded a notebook and try to launch
Martin Wolstencroft: I can answer you directly that the new box titles that we merged yesterday broke the notebooks in a way that I didn't notice and wasn't caught (despite linting expliticly for that.)
It should be fixed this afternoon, as of the merge of the abcd-fix branch
please check again now
I believe it is fixed
OK, I did wonder because I noticed that some of teh lessons had been updated very recently, checking now
Yep! I added a couple new lessons just now (venvs, multiprocessing, conda) and noticed the issue only today prepping for my class.
Martin Wolstencroft: if you find the notebooks useful, or needing changes, let me know. Until now I believe I've been the primary consumer of them.
The CLI basics notebook loads, I'll test another one at random... I'll let you know, at the moment I'm checking out which lessons we might put together into a CLI/git/Python/R introduction course to complement the HPC training I currently deliver
2 replies
Python advanced (wget https://training.galaxyproject.org/training-material/topics/data-science/tutorials/python-advanced-np-pd/data-science-python-advanced-np-pd.ipynb) works as well
thank you!