HP-UX 11.31 core file settings lost

[PamelaJThrasher]

Hello-
A while ago I had used coreadm on my HP-UX 11.31 servers to globally create any core files in /var/adm/crash.

When a core file filled up /home on one of these servers last night, I re-checked the settings with coreadm and found them to be back to the default.

I know the servers have all been rebooted since I initially created the core file settings. I suspect we lost them at that time but cannot be 100% sure.

This is the command I ran:
coreadm -e global -g /var/adm/crash/core.%p.%u.%f.%t

I want these settings to persist after the server is rebooted. Do I need to do something else?

Thanks in advance for your help.

[Steven E. Protter]

Shalom,

coreadm seems to be a new feature for 11.31 HP-UX.

http://docs.hp.com/en/B2355-60130/coreadm.1M.html

I have checked and there are no patches for this product.

If you used it correctly what happened should not have happened.

Still issues:
1) product fault
2) user fault

Hopefully the man page will be helpful. If there are indeed permanent settings for this tool (I'm not sure) and it acts at a daemon there should be configuration in /etc/rc.config.d/

I'm checking one of my 11.31 boxes and will add later if necessary.

SEP

[Tim Nelson]

It does state in the man page that the global settings are perm across reboots.

I guess if this is untrue then a possible workaround would be to create a boot script with your above command to execute at boot until a patch/fix is available. I would log a call so it gets acknowledged.

Did you verify that your settings have changed ? Perhaps the application or user's profile disables your changes ?

[melvyn burnard · Accepted solution]

Well this IS a new feature a of HP-UX 11iv3, and any chages done by the super user should be retained across reboots. I have just checked on one of my servers, and it had nothing set, then I applied the change you have specified, rebooted my box and it comes back:
# coreadm
global core file pattern: /var/adm/crash/core.%p.%u.%f.%t
init(1M) core file pattern:
global core dumps: disabled
per-process core dumps: enabled
global setid core dumps: disabled
per-process setid core dumps: disabled

So it DID persist across a reboot


I would check to see that the change did indeed take effect BEFORE doing a reboot, or see if this is not happening for only a spcific user who mauy have overridden this setting locally

[PamelaJThrasher]

Thank you to everyone that responded.

I do believe that I used the command correctly and had confirmed that it was working at one time.

I will just have to make a note to check the settings periodically especially before and after reboots.

Thanks again.

[PamelaJThrasher]

The problem is not really resolved but I will keep an eye one things per the suggestions from the responders.

[Dennis Handly]

Scanning the kernel sources indicates the coreadm info is stored with krs(5), most likely in: /stand/krs/