02-18-2014 08:05 AM
Hi
I am getting alerts showing that I am out of disk space in the /var directory on my LMS 4.2.1 server. The offending file seems to be a
tmp_000000 in /var/adm/CSCOpx/files/.SQLAnywhere/upmdbengine/tmp which is > 50Gb in size. Can anyone tell me :
a) why has this file got to this size (and how do I stop it happening again)
&
b) Can I safely delete this file.
Thanks
Solved! Go to Solution.
02-18-2014 08:28 AM
The corresponding temporary file is created on database server startup and it will be deleted on shutdown. The file will hold the query transactions that happened with the corresponding database while the database is running and used when there is not enough cache memory available for executing the queries.
You can delete this file after stopping LMS daemon.
Most probably the reason behind this would be UPM is polling more MIB objects than supported on your server, which is doing a lot of transactions on DB.
As a try to address this issue, you can try to re-initialize the UPM database, which resets the DB to like a fresh installed server. All custom settings like templates and pollers will be removed and UPM/Performace module will be like Factory reset.
If not details are needed from it, you can re-initialize using following commands :
Stop the LMS severices :
/etc/init.d/dmgt stop
/opt/CSCOpx/bin/dbRestoreOrig.pl dsn=upm dmprefix=UPM
Start the service:
/etc/init.d/dmgt start
The root cause of this file growing is unknown yet and have not been seen on many installtions. It is very rare. Open a TAC case to get it troubleshooted as LMS Development team would be involved to troubleshoot Database transaction file. But mostly their one of the first try is the re-init of DB.
-Thanks
Vinod
**Rating Encourages contributors, and its really free. **
02-18-2014 08:28 AM
The corresponding temporary file is created on database server startup and it will be deleted on shutdown. The file will hold the query transactions that happened with the corresponding database while the database is running and used when there is not enough cache memory available for executing the queries.
You can delete this file after stopping LMS daemon.
Most probably the reason behind this would be UPM is polling more MIB objects than supported on your server, which is doing a lot of transactions on DB.
As a try to address this issue, you can try to re-initialize the UPM database, which resets the DB to like a fresh installed server. All custom settings like templates and pollers will be removed and UPM/Performace module will be like Factory reset.
If not details are needed from it, you can re-initialize using following commands :
Stop the LMS severices :
/etc/init.d/dmgt stop
/opt/CSCOpx/bin/dbRestoreOrig.pl dsn=upm dmprefix=UPM
Start the service:
/etc/init.d/dmgt start
The root cause of this file growing is unknown yet and have not been seen on many installtions. It is very rare. Open a TAC case to get it troubleshooted as LMS Development team would be involved to troubleshoot Database transaction file. But mostly their one of the first try is the re-init of DB.
-Thanks
Vinod
**Rating Encourages contributors, and its really free. **
02-19-2014 04:06 AM
Thanks Vinod. Stopped the services and deleted the file although that didn't free the space and had to use lsof to track down the process that had the file locked open and kill that before I got the disk space back. Will keep an eye on it and if it re-occurs will follow your other suggestions in order.
Regards
Alan
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide