Would be interesting to find out which one of those caused your issue... but not really important :p My guess is the first change you made. Perhaps you should have mentioned you are using qsub to submit the job, I think that was important information.
Hi all,
I have a STAR alignment job that has been running for 2 days. Here is command:
STAR --genomeDir GENOME_DIR --readFilesIn $FASTAQ_DIR/read_1.fastq $FASTAQ_DIR/read_2.fastq --runThreadN 16
And in the end of Log.out file, it just shows "Create thread #..." No error massage and any complaints were found. Any ideas?
Thank you in advance!
1 answer
Hello,
I finally make it work! Here are two changes I made:
Specify qsub option -l m_mem_free=8G. With multiple cores, more than 30G memory is available.
At first, I used the STAR installed on the server. genomax2 reminds me that it is outdated. I am not sure this is the key issue, but I download the latest version STAR-2.5.2. Now everything is fine.
Thank you all!
Log in to answer this question.
If nothing has happened for 2 days this is not going anywhere. You would be best off killing the job and restarting. Are you trying to use more cores than are available to you/on your server?
Thank you for reply.
I have run this job twice. For the second time, I used option -pe as 16, but it still got me nowhere. I found a similar post on google groups. They seemed to have no answer. I killed the job anyway. similar post
How much memory do you have available on this server? STAR needs a minimum of 30G.
Are you using pre-compiled STAR or did you compile the program yourself? Are you using the the latest version?
Can you see the job running/using memory or is it completely dead?
I also spot a likely error in your command: you specify GENOME_DIR but that should probably be the environmental variable $GENOME_DIR
Thank you for pointing it out. That was just a typo. I use $GENOME_DIR in my command. I check the job status. It shows running, but hangs on the 'create thread' step Still no idea why that happened...
Have you tried with just one thread?