Hi
I'm seeing something a little strange at the moment.
Let's say that I've just started the server process and check the syslogs that have been captured, they all look fine and the timestamps are 100% right. Now, give this a few days and the syslogs start showing completely different times. It's almost like there is a cache of logs that are being released into the database slower. Obviously, the timestamp in the syslog message itself is correct, as I've verified this on all the devices I've seen this for. This does only seem to happen for devices that have a large volume of messages towards NetXMS.
I've attached a screenshot to show what I mean.
BTW, the 8:59AM in the screenshot was correct, the latest logs I had in there at that time, though it shows logs from yesterday evening.
Any help would be appreciated, it seems that the only resolve for this right now is to periodically restart the server process.
Thanks
Hi!
Similar problem was reported on Russian forum as well. It seems that NetXMS syslog server cannot write incoming messages to database fast enough. I'll try to do load test in my lab and hopefully fix that.
Best regards,
Victor
Thanks Victor!
Quote from: Victor Kirhenshtein on February 19, 2014, 11:59:42 AM
Hi!
Similar problem was reported on Russian forum as well. It seems that NetXMS syslog server cannot write incoming messages to database fast enough. I'll try to do load test in my lab and hopefully fix that.
Best regards,
Victor
Hi Victor,
Just another bit for you to know here... I migrated away from SQL Server and went to MySQL. I get the same issue with this too.
Thanks
Quote from: fj89j on February 19, 2014, 12:07:52 PM
Thanks Victor!
Hi Victor
I noticed that when this issue seems to rear its head, all messages appear to be cached in RAM. Things can be fine for quite a while, but as soon as the process takes so much RAM that the server cannot operate as it should, it causes crashes
Thanks