Video Screencast Help

Custom format of logs in engine_A.log

Created: 14 May 2013 • Updated: 19 Jun 2013 | 4 comments
This issue has been solved. See solution.

Hey

Do you happen to know if there is a way to tweak VCS in a way that it carries additional information in every log entry? Preferebly, I would like for every entry in the log to display cluster name and id, before or after the error code. Why? I need that for Tivoli monitoring to be able to identify same events coming from the same cluster and as a result creating one alert and not as many as the number of cluster nodes. Only by adding those two I would feel comfortable enabling deduplication on the tivoli end. Together with event summary they would create a unique key. If there is no way to do so I cannot enable it due to the size of the environment that can have similar res/group names sharing certain infrastracture componenta such as SAN or network of which issue can result in failures on a number of clusters at the very same time.

Thanks
Wojtek

Operating Systems:
Discussion Filed Under:

Comments 4 CommentsJump to latest comment

arangari's picture

Hi Wojtek,

The "LogClusterUUID" attribute at cluster level will log the  clusters UUID  stored in attribute 'CID' for every log. The clustername may not be unique, however the cluster UUID is going to be unique and we can leverage it. 

Let me know if this helps in resolving the issue for you. 

Just to query - are you 'reading' engine_A.log file for some post-processing? If you could provide bit more details with example on what you are looking for, it will be useful to suggest better solution.

Thanks and Warm Regards,

Amit Rangari

If this post helped you resolving the issue, please mark it as solution. _____________________________________________________________________________

SOLUTION
arangari's picture

You still didnt provide the reason you are reading enigne_A.log.

 

Thanks and Warm Regards,

Amit Rangari

If this post helped you resolving the issue, please mark it as solution. _____________________________________________________________________________

bazi's picture

Appologies for late reply and thanks for the follow up. What we do is we use Tivoli log adapters to monitor cluster for undesired events. With properly build regex's we are able to differ alerts to dfferent support groups which use different ticketing groups in our incident tracking systems, by filtering them by certain keywords in the names of resources and groups. We follow very strict naming convention that allows us to do it.

In one of your posts you mentioned I could use cluster uuid. If I am not mistaken this feature was introduced in SF5.x. If I am correct that would be a problem as we still have some older build v4.x awaiting migrations. Neverthless I do not see the id appearing in the log. Is it an option I need to enable?

Thanks & regards
Wojciech

arangari's picture

@Bazi:

The cluster uuid feature was introduced in VCS5.1 (SFHA 5.1). For older versions you may not be able to use this method.

yes - this needs to be enabled  - check the value of cluster attribute LogUUID. 

however i feel this will help you to segregate the messages from same 'cluster'. 

but you have different ticketing groups etc - where the groups names or resource names will be used. so some kind of regex is must have. 

have you tried 'hareg -cache' ?

 

Thanks and Warm Regards,

Amit Rangari

If this post helped you resolving the issue, please mark it as solution. _____________________________________________________________________________