Linux Agent - Will not run as 1 user with error code 11

[Tina Hart]

Good Morning,

I wanted to know if  anyone has ever ran in to this issue with a Linux agent.  We have it configured so that the main binary runs as root so that multiple users can run jobs using this Agent.  This works properly, however, 1 user can run a few jobs successfully and then we start to receive the following errors (i have no idea why the error comes back in French):

U02003043 Impossible de recevoir du partenaire 'execle()'. Code erreur '11'.

We are looking at potential issues with the user account in this case, but our team is not able to find any details. The strange thing is if I restart the Agent, the jobs will work 1 time, and then start to fail, so not sure if we're missing anything.

The other users that run jobs on this agent are always working (there are 3 other accounts that use this agent).

Has anyone ever come across this or have any advice on what we can do?
Thanks
Tina

[Frank Muffke]

Hi Tina

regarding french => Agent inifile: language=(E,D)

which linux do you use, which env? pls show a listing of the binaries....

cheers, Wolfgang

------------------------------
Support Info:
if you are using one of the latest version of UC4 / AWA / One Automation please get in contact with Support to open a ticket.
Otherwise update/upgrade your system and check if the problem still exists.
------------------------------

[Tina Hart]

Hello,

So the Agent file language has E and D only (no French, so very strange).  Sorry, I forgot to send which versions we're actually on.  Our Linux agent is 12.3.1 and this is an agent that's actually on the Oracle DB server.  Other user accounts can run jobs on that server, and just not this one.  It was recommended that we try to upgrade the agent, but as there are other jobs relying on the agent that are working, it's hard to complete the upgrade at this time.
Strangely enough, it looks like after there was a scheduled restart of that server (or if restarting the agent), the job ran 1 time successfully and then started to fail again.  

Here is a list of the binaries (accounts/group other than root change for security :) )

-rwxrwxr-x 1 autouser ogroup   78710 Dec 21  2019 ucxelx6f

-rwsrwxr-x 1 root    ogroup 3607125 Dec 21  2019 ucxjlx6

-rwxrwxr-x 1 autouser ogroup   19837 Nov 29 09:07 ucxjlx6.ini

-r-------- 1 root    ogroup      83 Mar 20  2020 ucxjlx6.kstr

-rwxrwxr-x 1 autouser ogroup  262728 Dec 21  2019 ucxjlx6m

-rwxrwxr-x 1 autouser ogroup   80255 Dec 21  2019 ucxjlx6s

-rwxrwxr-x 1 autouser ogroup   19814 Dec 19  2019 ucxjxxx.ori.ini

-rwxrwxr-x 1 autouser ogroup 2954865 Dec 19  2019 ucx.msl

Right now we are testing this out in Development, so the process has not moved to production.  Our DBAs are trying to run their backup scripts from Automic, which work 1 time and then fail.  The same also happens with a basic job that displays env details.  It also looks like each time, the error can vary.  On the run last night, I saw the following in the Agent log file:
U02000009 Job ended with return code '210'

I'm starting to wonder if there is something wrong with that specific user id, but if there's anything you've noticed before regarding this it would be really great.

Thanks for your help!
Tina

[Jared Kessans]

We had a similar issue in the past where we could run jobs on a Linux agent but one user was getting the following error.
"U2003043 Invalid 'setuid()' call. Error code: ('11' - 'Resource temporarily unavailable')"

The jobs would abend with a return code 203.

It was all related to the ulimit.  While I am not a Linux admin so I had to rely on our Hosting group for a fix, I never got a straight detailed fix but I believe the information below was related.

They updated the nproc –n and –u to 2048 each on the agent's ID and possibly the ID used to run the job.  I have a coworker that states that they were upped even higher to 16384.  Again I never got a detailed explanation from our Linux admins even though we have seen this a couple of times over the years.

[Devidutta Das]

Yes we had a similar issue and UNIX admin updated the limit of  nproc.conf file to 65536 and it fixed the issue.I think the default is 1024.

[Tina Hart]

Thanks Everyone!

The team has decided not to pursue the jobs they were trying to run with this user at this time, and no further issues have come up.  I will keep these notes for our admins as we may need to review this in the future.