CA TUESDAY TIP: IT'S ALL ABOUT PERFORMANCE.... AGAIN

[Sandra_Antunes]

Hey everyone

This week's topic is about Performance again, but with a different approach. I recently had a customer asking what could be done to avoid performance issues in Service Desk Manager.

We all know there are many, really many factors that can cause performance issues to an environment. However, there are some "best practices" that help keeping your environment "healthier" :grin:

Here are some tips:
[list][*]Run periodic archive and purge on session log and activity notification tables.
They can grow very quickly. If they are already highly populated, you should not do it all at once, but schedule short archive periods until you are able to have them small enough to run archives/purges once a month, for example.

[*]Depending on the MDB size, have periodic maintenance done to it, such as reorganizing tables and indexes. If running SQL Server, don't forget to have the transaction log file shrinked.

[*]Avoid having contacts with duplicate userids on ca_contact table.

The following query can be used to find them:

SELECT userid,
COUNT(userid) AS NumOccurrences
FROM ca_contact
GROUP BY userid
HAVING ( COUNT(userid) > 1 )

You run it and found some. Now what?
Never, ever delete them. You can cause inconsistencies to your environment.
If it is the same user, you can deactivate one of them.
Now, no matter if the user is the same or not, the userid should be changed. Even having the contact deactivated doesn't mean it won't be found.

[*]Running BOXI to generate your reports? Consider running "offiline reports" if you are having performance issues.
The CA_Business_Intelligence_for_CAService_Desk_Manager Greenbook explains how to configure it.

[*]Avoid having wrong email addresses set to users.

[*]Avoid having a notification method set without an email;

[*]The user must disconnect from SDM, not simply close the browser in the X. There are no warranties that the session will be killed by SDM.

[*]Don't leave traces active without need. You may have one active in your environment and doesn't even know it.
One simple way to check it is to run the following on the prompt:
pdm_logstat -vL

It should return an blank line.
If not and you don't remember of having this activated by any reason, you should contact CA Support to deactivate it.

Note there are other types of traces that can be used. If you suspect there are any active, you should contact CA Support to check if you have them active.

[*]You just upgraded to SDM 12.7? Check Monitor Joins. It's certainly active.
Consult TEC593935 - Critical Option to configure in Options Manager: Monitor Joins to understand its effects on SDM (guess what, performance issue... :tongue ), and in case you are really facing it, set this variable to NO.
[list]

Don't forget also to check other CA Tuesday Tips already published which talk about performance.

This is all for today....

Have you all a great week!
:wink:

Sandra Antunes

[Kyle_R]

Sandra_Antunes wrote:

This week's topic is about Performance again, but with a different approach. I recently had a customer asking what could be done to avoid performance issues in Service Desk Manager.

We all know there are many, really many factors that can cause performance issues to an environment. However, there are some "best practices" that help keeping your environment "healthier"

Thank you, Sandra,

Performance is a topic close to my heart and this is a fantastic article on items to "check off" as part of a health check for an SDM environment.

Every site should run a health check at intervals, such as annually. A checklist such as this one really helps.

Thanks, Kyle_R.

[Anon Anon]

[size=5][color=#5116b6]
Thanks for you extra performance tips I will add them to my maintenance routines.

Here are some of my mine-
I set up a permanent archive job in Archive and Purge for Auditlog, Event_Log, Knowledge Search Log,
Notify and Session logs -- It run to purge early Sunday mornings.

I also reindex the knowledgebase on Sundays as well -- pdm_k_reindex
Knowledge is much happier since I've been doing this.

I also create a new default attachment Repository at the beginning of each year as
part of the incident/request/problem/change purge archive activity (everything closed
and older than 2 years is archived).
cheers
Donna
[color] [size]

[Sandra_Antunes]

vorlon15 wrote:

[size=5][color=#5116b6]
Thanks for you extra performance tips I will add them to my maintenance routines.

Here are some of my mine-
I set up a permanent archive job in Archive and Purge for Auditlog, Event_Log, Knowledge Search Log,
Notify and Session logs -- It run to purge early Sunday mornings.

I also reindex the knowledgebase on Sundays as well -- pdm_k_reindex
Knowledge is much happier since I've been doing this.

I also create a new default attachment Repository at the beginning of each year as
part of the incident/request/problem/change purge archive activity (everything closed
and older than 2 years is archived).
cheers
Donna
[color] [size]

Thanks for adding more items to the checklist, Donna.
Any extra input is more than welcome.

Cheers,
Sandra

[System]

Thanks for all the great info Sandra! Thanks also for the added info Kyle and Donna:grin:

[wteixeira]

Thanks for all those nice tips.