[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[LDM #FKC-383851]: rtstats



Austin,

> I've got an LDM installation here at IARC that has been working pretty
> well since December. Last night there was a problem with the machine
> requiring a restart, and after starting up ldm, there have been some
> problems. No new data seems to have been saved and rtstats has logged
> near 100% CPU, adding up to hours and hours of CPU time. I've tried
> restarting ldm (which the first time required manual kill -9 and clean
> up), but the same problems reappear after a few minutes.
> 
> I've attached the ldmd.conf and the latest ldmd.log.

The LDM log-file contained this

...
> Aug 19 13:09:40 dixon pqact[3084] NOTE: Starting Up
...
> Aug 19 13:09:40 dixon pqact[3084] NOTE: Starting from insertion-time 
> 2008-04-09 19:42:43.880883 UTC

It looks like the crash left the .state file that pqact(1) uses
in a bad state (2008-04-09 is too long ago).  The pqact(1) process
then died:

> Aug 19 13:09:40 dixon pqact[3084] ERROR: pq_sequence: xdr_prod_info() failed
> Aug 19 13:09:40 dixon pqact[3084] ERROR: pq_sequence failed: Input/output 
> error (errno = 5)
> Aug 19 13:09:40 dixon pqact[3084] NOTE: Exiting

I suggest trying the following:

1.  Stop the LDM (ldmadmin stop).

2.  Remove all the .state files (rm $HOME/etc/*.state)

3.  Start the ldm (ldmadmin start).

4.  Check the LDM log-file for errors.

Regards,
Steve Emmerson

Ticket Details
===================
Ticket ID: FKC-383851
Department: Support LDM
Priority: Normal
Status: Closed