WebMO - Computational chemistry on the WWW
Recent news

WebMO 17.0 is now available for free download!

WebMO 17.0 Pro and Enterprise have a variety of additional features and is available for purchase.

The WebMO app is now available for iOS and Android.

August 23, 2019

Submitting jobs on Rocks 6.2 Log Out | Topics | Search
Moderators | Edit Profile

WebMO Support Forum » WebMO Installation » Other » Submitting jobs on Rocks 6.2 « Previous Next »

Author Message
Gary Hoffman
Unregistered guest
Posted on Friday, August 19, 2016 - 10:03 am:   

I just installed Rocks 6.2 on a cluster and have tried to set up WebMO Enterprise, v. 15.0. With the WebMO queueing system, I can run jobs just fine (but they run on the headnode). I set up the SGE queueing system as described in the instructions. A submitted job generates pbs_script.sh and then the job ends with the message "Cannot open file /home/webmo/webmo/ggh/9/run_log: No such file or directory at text_dump.cgi line 87". The job is not submitted to SGE. If I submit pbs_script.sh at the command line as user webmo using the command specified in the script, it runs fine. The ports for sge_qmaster and sge_execd are included in /etc/services and sge_root is generated just fine. Based on one entry in this forum I added the line 'Defaults env_keep += "SGE_CELL SGE_ARCH SGE_ROOT"' to /etc/sudoers, but this makes no difference. It seems I just need to get qsub to execute, but I don't know how to get there. Can you help?
JR Schmidt
Moderator
Username: Schmidt

Post Number: 517
Registered: 11-2006
Posted on Friday, August 19, 2016 - 10:16 am:   

Gary,

1) Check the paths to 'qsub' in the 'Batch Queue Manager'
2) Check the WebMO error log (/home/webmo/webmo/errors) for more details.
Gary Hoffman
Unregistered guest
Posted on Friday, August 19, 2016 - 11:56 am:   

JR,
Thanks for the quick response. I double checked the path for qsub (and the other commands). They look correct. In the error log (why didn't I think of that before?) I found a lot of errors of the form:
"error: commlib error: got select error (Connection refused)
error: unable to send message to qmaster using port 6444 on host "parcore.local": got send error"
The port for sge_qmaster is listed as 6444 in the /etc/services file. Does this make sense to you?
JR Schmidt
Moderator
Username: Schmidt

Post Number: 518
Registered: 11-2006
Posted on Friday, August 19, 2016 - 12:03 pm:   

Are you sure the port number is correct? Based on the "connection refused", either the port is incorrect or some firewall rule is blocking the connection.
Gary Hoffman
Unregistered guest
Posted on Friday, August 19, 2016 - 2:07 pm:   

Pardon my ignorance in these matters; I feel I am over my head. I had relied on /etc/services to get the port number for sge_qmaster. Using netstat, I got a port number of 536. I don't know if that port number will persist or where I could find the number with the script if it changes. Anyway, I hardwired the number in processcontrol_pbs.cgi and it seems to be running now. Thanks for your help.
JR Schmidt
Moderator
Username: Schmidt

Post Number: 519
Registered: 11-2006
Posted on Friday, August 19, 2016 - 2:16 pm:   

Sounds like SGE on is not configured to use the default port (or it changed) and /etc/services was never updated to reflect this. This is probably a bug with Rocks (or some underlying distribution), technically.

Add Your Message Here
Post:
Username: Posting Information:
This is a public posting area. Enter your username and password if you have an account. Otherwise, enter your full name as your username and leave the password blank. Your e-mail address is optional.
Password:
E-mail:
Options: Post as "Anonymous"
Enable HTML code in message
Automatically activate URLs in message
Action:

Topics | Last Day | Last Week | Tree View | Search | Help/Instructions | Program Credits Administration