Huge elog* Files

[David Teh]

Hi folks,

Cloverleaf 5.6 on Solaris.

Due to a situation which have been rectified, we end up having huge elog* files in the ../exec/databases directory.

Any way to reduce the sizes with the site up and running, and without destroying any data?

TIA!

Topic

[Robert Milfajt]

My understanding, and anyone is free to correct me if I’m wrong, if the recovery database (and error database) are empty, there is not data to destroy.  If you had a huge backlog of messages, Raima grabs the space and doesn’t release it, so you need to initialize your databases.  A typical procedure would be to stop all inbound threads, make sure all messages are out of the recovery database (hcidbdump -r returns null).  Probably the same for error database (hcidbdump -e).  Then stop all outbound threads, then stop the daemons, then initialize the database (hcidbinit -AC), then start things back up in reverse order.  But I would defer to and probably call Lawson support before doing this to make sure.

Hope this helps,

Robert Milfajt
Northwestern Medicine
Chicago, IL

[David Teh]

Hi Robert,

Thanks for that. That’s pretty much the standard procedure which I am trying to see if I can avoid. 🙂

Guess there’s no way to do it while everything is still up and running.

Robert Milfajt wrote:

My understanding, and anyone is free to correct me if I’m wrong, if the recovery database (and error database) are empty, there is not data to destroy.

[Author]

Replies