-
Notifications
You must be signed in to change notification settings - Fork 5
trouble
Slurm fails... This is apparently normal for Galaxy 18.05 and 18.09.
It could be an issue with your Resource Libraries. Make sure you have run all of the data manager tools from admin or from the setup-data script. If it's input files that aren't showing up, double check the datatype of the file. This can be found in the pencil icon on the top right of the entry in your history. Some tools are not forgiving about whether a fastq file is fastqillumina or fastqsanger, so check and make sure. It's a good idea to generate a tour of the tool to see an example of the tool working, and what kind of files it takes.
If docker messes up, it might be because there already was a docker running.
#Job for docker.service failed because the control process exited with error code. See "systemctl status docker.service" and "journalctl -xe" for details.
#Check the system journal for clues...
journalctl -xe
#Try removing a stale pid
rm /var/run/docker.pid
#Check status
systemctl status docker.service
# If needed, wipe out all docker processes
ps axf | grep docker | grep -v grep | awk '{print "kill -9 " $1}' | sudo shCan't kill the docker container:
docker stop fa29fb2feefa
Error response from daemon: cannot stop container: fa29fb2feefa: Cannot kill container fa29fb2feefac2f6dd1c24f2f5ba875b58da305366ed8fd6fda8870c2a3bf277: connection error: desc = "transport: dial unix /var/run/docker/containerd/docker-containerd.sock: connect: connection refused": unknown
#restart docker daemon if needed
systemctl restart dockerDocker uses something called Overlay in order to separate disk space inside the docker image from that available to the whole host. See the [development page]{dev} for more information about swapping out the normal docker disk for a larger native size, or if docker is consuming all of your disk space during the build process. DO NOT just delete files out of /var/lib/docker, it's a huge pain to fix. You usually have to run the commands from the dev page anyway.
If you are using the export command with docker run, and you are still running out of space, check out the folder /export/resources and delete any unnecessary tar.gz files. These were probably used to build your libraries, but once all is built and well, you won't need those files anymore.
STAR is ususally a pretty happy-go-lucky program, and does not usually cause issues. Howerver, there are a few circumstances to be careful of.
The most common cause of failure for STAR is memory. Generally, 40gb of memory should be set aside for running STAR against a human reference. Make sure your VM can support that.
During the data manager steps of setting up the Docker instance, you may get cryptic failures with some of the steps, or Galaxy itself may die. The first thing to check is the permissions in /export. If Galaxy can't write to /export, it won't be able to start. If Galaxy can't write to your Resource Library, your data manager may fail. Both of these symptoms can b e alleviated by running preflight.sh from your Host, or chowning the /export directory to Galaxy once inside your container.
The following errors can happen while you are building your resource libaries:
RuntimeError: Error, didn't locate single mutation_lib.*.tar.gz in /export/resource/ctat_lib/GRCh38_v27_CTAT_lib_Feb092018/ctat_genome_lib_build_dir
This can happen when more than one mutation tar file is present in the Resource Library directory. You should choose either hg19 or hg38, put them in separate folders, or at least label them differently.