cron: resource control assignment failed

cron: resource control assignment failed

Post by Va » Sat, 24 Jun 2006 18:51:52



Solaris 10 1/06 on SunFire 210

This is my first Oracle installation on Solaris 10.

I followed the recommendation and removed all shared memory
entries from /etc/system.

I added a default project for user oracle:

user.oracle:100::oracle::project.max-shm-memory=(priv,2147483648,deny)

User oracle has a bunch of cron jobs to do all the db maintenance stuff
like exports, rman backups and the like.

And here is the problem:

When cron tries to execute the scripts, I get errors like this:

Jun 23 03:00:00 host cron[18887]: [ID 832980 auth.error] pam_unix_cred: project.max-shm-memory resource control assignme
nt failed for project "user.oracle"

The strange thing is that in some cases the cron jobs run fine, so this error
seems to appear randomly. All the scripts executed by cron are similar
(they use sqlplus to do something on the db)

I digged into sunsolve and found one entry that may be related to this: Document ID 6194864
However there seems to be no solution for this.
Google couldn't help either.

Any ideas?

Van

 
 
 

cron: resource control assignment failed

Post by Stu » Sat, 24 Jun 2006 19:44:49



> Solaris 10 1/06 on SunFire 210

> This is my first Oracle installation on Solaris 10.

> I followed the recommendation and removed all shared memory
> entries from /etc/system.

> I added a default project for user oracle:

> user.oracle:100::oracle::project.max-shm-memory=(priv,2147483648,deny)

> User oracle has a bunch of cron jobs to do all the db maintenance stuff
> like exports, rman backups and the like.

> And here is the problem:

> When cron tries to execute the scripts, I get errors like this:

> Jun 23 03:00:00 host cron[18887]: [ID 832980 auth.error] pam_unix_cred: project.max-shm-memory resource control assignme
> nt failed for project "user.oracle"

> The strange thing is that in some cases the cron jobs run fine, so this error
> seems to appear randomly. All the scripts executed by cron are similar
> (they use sqlplus to do something on the db)

> I digged into sunsolve and found one entry that may be related to this: Document ID 6194864
> However there seems to be no solution for this.
> Google couldn't help either.

> Any ideas?

> Van

According to the opensolaris bug database it's a bug and being worked
on.  I have similar problems.  I have a 2 node VCS cluster with
multiple databases on each server.  I get a random VCS alert sometimes
that hard oracle monitoring has failed and in the VCS engine log an
error stating the erorr message you get above which is also logged to
system messages.

If you kick off your backups at once, I found running them in
sequential order worked around it.  Also, try staggering your cron
jobs.

Stu.

 
 
 

cron: resource control assignment failed

Post by Va » Mon, 26 Jun 2006 07:09:20



> According to the opensolaris bug database it's a bug and being worked
> on.  I have similar problems.  I have a 2 node VCS cluster with
> multiple databases on each server.  I get a random VCS alert sometimes
> that hard oracle monitoring has failed and in the VCS engine log an
> error stating the erorr message you get above which is also logged to
> system messages.

> If you kick off your backups at once, I found running them in
> sequential order worked around it.  Also, try staggering your cron
> jobs.

> Stu.

Thanks for the info.
In this specific case, staggering the cron jobs didn't eliminate the errors.
However I found a workaround that works:
I now have one single cron job which runs shortly after midnight and
that creates all the scripts for the day to be run as at jobs.
Works fine so far.

Van