View previous topic :: View next topic |
Author |
Message |
jbarrellon
Joined: 08 Oct 2008 Posts: 23
|
Posted: Mon Feb 17, 2014 4:25 am Post subject: Logging pool is full |
|
|
Hello,
More and more often, our HM primary logging method (Oracle database) stops working and we're getting following messages in HM html log file :
02/14/2014 8:01:46 Monitoring service warning: Logging pool is full
02/14/2014 8:03:10 Warning: Logging performance is very low considering average test frequency
After restarting HM, it's working again.
HM performs about 4 tests/s (i would say about 7-8 tests/s during business hours).
Please can you explain in what case exactly you display this kind of message in log file ?
I'd like to know if there is something we can do about this or if we will have to change our default logging mode (currently Full). I'd prefere to avoid that because there will be a huge impact for us.
HM v9.58 running on Windows Server 2008 R2 Standard.
Thank you
Julien |
|
Back to top |
|
|
KS-Soft
Joined: 03 Apr 2002 Posts: 12795 Location: USA
|
Posted: Mon Feb 17, 2014 5:04 am Post subject: |
|
|
Quote: | More and more often, our HM primary logging method (Oracle database) stops working |
Please do not use Oracle software (especially Oracle ODBC drivers version 10.xx), there are hundreds bugs!
If you must use Oracle database, try Microsoft ODBC driver for Oracle or Oracle client version 9.2.0.8.0 - as we know this is the most stable version.
Version 11.xx better than version 10.xx but still has some problems.
Quote: | 02/14/2014 8:03:10 Warning: Logging performance is very low considering average test frequency
Please can you explain in what case exactly you display this kind of message in log file ? |
This means logging too slow.
HostMonitor already got new test results while old results are not stored to Oracle database yet
(yes, there is internal HostMonitor buffer but its full of data that was not stored yet)
Quote: | HM performs about 4 tests/s (i would say about 7-8 tests/s during business hours) |
Its not a problem at all for any "normall" database.
Oracle ODBC driver may work fine for a while and then cause memory or handles leakage, high CPU Usage, access violation errors and so on. It can crash HostMonitor easily and slow down entire system.
Regards
Alex |
|
Back to top |
|
|
jbarrellon
Joined: 08 Oct 2008 Posts: 23
|
Posted: Thu Feb 20, 2014 4:00 am Post subject: |
|
|
Quote: | try Microsoft ODBC driver for Oracle or Oracle client version 9.2.0.8.0 |
We're using Oracle client version 11.2.0.3 and it won't be possible for us to downgrade to 9.2.0.8.
We changed from Oracle ODBC driver to Microsoft ODBC driver and it seemed to be a little bit better at the beginning but actually we still have issues.
Quote: | there is internal HostMonitor buffer but its full of data that was not stored yet |
Is there a way to increase the size of HostMonitor internal test buffer ?
Also, we're regularly having the following messages :
More than 8 consecutive log errors occured. Primary logging paused for 5 min (use Options dialog to resume logging manually).
We're currently analysing to resolve the problem causing this, but meanwhile i cannot see the resume option in Options dialog box. Where is it ? And even better, is there a way to disable this behavior because we cannot accept losing logs.
Thank you
Julien |
|
Back to top |
|
|
KS-Soft
Joined: 03 Apr 2002 Posts: 12795 Location: USA
|
Posted: Thu Feb 20, 2014 9:26 am Post subject: |
|
|
Quote: | Is there a way to increase the size of HostMonitor internal test buffer ? |
If ODBC driver does not work, how this can help?
What sense to increase buffer if logging does not work anyway?
Quote: | More than 8 consecutive log errors occured. Primary logging paused for 5 min (use Options dialog to resume logging manually).
|
Please check system log, you will see error description there.
Regards
Alex |
|
Back to top |
|
|
KS-Soft
Joined: 03 Apr 2002 Posts: 12795 Location: USA
|
Posted: Thu Feb 20, 2014 9:32 am Post subject: |
|
|
Quote: | but meanwhile i cannot see the resume option in Options dialog box. Where is it |
Simply open Options dialog and click Save button.
But again - it does not make any sense until you fix the problem.
1st you should check system log, what EXACTLY error do you see?
May be you have wrong query that cannot work with some data (e.g. you are trying to store text into numeric field).
Quote: | And even better, is there a way to disable this behavior because we cannot accept losing logs. |
Sure, just switch to file logging.
Then check your logs, find what exactly is wrong with ODBC, call your network administrator and fix the problem.
Do you need ODBC logging? Do you need Oracle? May be you can use MS SQL Express?
May be network connection to SQL server too slow?
Regards
Alex |
|
Back to top |
|
|
jbarrellon
Joined: 08 Oct 2008 Posts: 23
|
Posted: Mon Mar 03, 2014 9:32 am Post subject: |
|
|
Quote: | Simply open Options dialog and click Save button. |
OK, thank you for the tip.
Quote: | But again - it does not make any sense until you fix the problem.
1st you should check system log, what EXACTLY error do you see? |
As i said, we're already working to solve the error in system log.
(Just for information :
03/03/2014 9:56:48 Primary log error. ODBC error code: 4294967295 ([Microsoft][ODBC driver for Oracle][Oracle]ORA-21561: OID generation failed ORA-06512: at "HM.TR_LOGHM", line 2 ORA-04088: error during execution of trigger 'HM.TR_LOGHM) Logging initiated by "WCS-DBLTK EventLog Application Warning" test. )
Quote: | Sure, just switch to file logging. |
We've already configured secondary log to file logging. Why does HM stop logging during 5 minutes after 8 consecutive errors instead of logging in secondary log ?
Regards
Julien |
|
Back to top |
|
|
KS-Soft
Joined: 03 Apr 2002 Posts: 12795 Location: USA
|
Posted: Mon Mar 03, 2014 9:51 am Post subject: |
|
|
Quote: | We've already configured secondary log to file logging. Why does HM stop logging during 5 minutes after 8 consecutive errors instead of logging in secondary log ? |
did you check backup log? no records? |
|
Back to top |
|
|
jbarrellon
Joined: 08 Oct 2008 Posts: 23
|
Posted: Fri Mar 07, 2014 1:40 am Post subject: |
|
|
For now, a workaround that consists in rebooting the server every night has been set up.
If the problem appears again, we'll check if there are records in backup log during the 5 minutes pause and let you know. |
|
Back to top |
|
|
|