HP SIM problem

[joseph pareti]

Apology for the lack of detail, currently this is all I know about this case. is there anything I should be aware of prior to the customer's meeting?

During the (SIM) installation no data base was created in our SQL setup.

We used a workaround, specifically the SQL-Express (included w/ SIM) was installed, and the created database was moved.  

During operation, at some points in time we have data loss, i.e. the discovered devices disappear. However the data are still in SQL, therefore we suspect some problems in connecting to the SQL server.

<jp note> Apparently this was not caused by firewalls. <jp note>

P.S. This thread has been moved from ProLiant Servers (ML,DL,SL) to ITRC HP Systems Insight Manager Forum - HP Forums Moderator

[Matti_Kurkela]

"...the created database was moved." You'll definitely want to know where it was moved to, and who maintains that system. (There are two applicable maintainer roles: a sysadmin and a DBA, and they may or may not be the same person.)

You should be able to get some clues by reading the database connection parameters from the SIM configuration if there is no other documentation about the move.

Then you want to know the history of planned and unplanned service breaks of that database system. Any correlation to the times of your data loss? Is the system maintained in an appropriate way to support your SIM? Also, is the database server overloaded by other databases?

If you're using SIM as an important management tool and the SIM database has been placed to a non-production "database test and general DBA playground" server which may be rebooted at any time without prior notice, you may want to have the database moved to a database server that is more appropriate for a production system.

You'd also like to know the rough location of the database server, in terms of network topology. Is the server:

• in the same network segment as the system that runs the SIM? (Good to know: History of planned & unplanned layer-2 network service breaks in that segment?)
• in a different network segment, but in the same datacenter? (...in both segments? History of service breaks in the layer-3 network equipment (i.e. router, firewalls or similar) connecting the two segments?)
• in a different datacenter altogether? (All of the above, plus has there been any known third-party backhoe events or other major issues on the link(s) between the datacenters?)

How does the network connection between the SIM server and the database server behave? Can you get any network load statistics?

If there are no other information sources, you might want to ping the database server from the SIM server for a long time (e.g. 24h) and look at the round-trip time statistics. If the average Round-Trip-Time is small and the difference between the minimum and maximum RTT is not significant, the basic network layer seems healthy. A large maximum RTT might indicate that some part of the network path is overloaded at some time(s) of day, which might cause database connection issues.

[joseph pareti]

we gathered the following new insights:

1. The HP SIM discovery process provides elusive and non repeatable patterns, i.e. out of 1200 systems it discovers sometimes just a few, today it discovered  400 odd,…
2. The initial install involved SQL EXPRESS as a workaround, the data was then migrated to MS SQL for productive use. Not sure how much this process follows the “HP rules”, especially comparing with http://h30499.www3.hp.com/t5/ITRC-HP-Systems-Insight-Manager/Migrating-HP-SIM-Database-into-new-SQL-Server/td-p/4759571
3. In the  mxdomainmgr.0.log log file there are repeated  javax.naming.NameNotFoundException  entries. FWIW I had similar problems in a totally different project (i.e. NOT HPSIM) and they were related to JNDI configuration errors, misplaced deployment descriptors, etc.

This leads me to think there is some kind of naming corruption, possibly by incorrect data migration, so that I would be inclined to recommend the following approach - out of the link posted above. Any comments?

STOP the HP SIM Service:-


1) go to C:\program files\HP\Systems Insight\Manager\config
2) edit database.admin
3) change hp.Database.hostName=
4) change hp.Database.instance=
5) edit database.props
6) change hp.Database.hostName=
7) change hp.Database.instance=
8) Edit *-ds.xml with new hostname and instance (in\jboss\server\hpsim\deploy)
-and-
there may be another file with a name like "tmpXXXXXpmp-ds.xml" in the folder \jboss\server\hpsim\tmp\deploy\ that you also should update.
10) START the HP SIM Service
11) wait for port 50000 to open
12) test.

[joseph pareti]

here is one example of the java error messages described before; they are all related to message driven beans and naming, i.e. JNDI

javax.naming.NameNotFoundException: MDBIPMIPowerCycle not bound

16 Oct 14:32:36,395 WARN [JMSContainerInvoker] creating a new temporary destination: queue/MDBIPMIPowerCycle

[shocko]

Check the following:

1. Community trap name used across systems
2. WBEM credentials
3. post identification logs

[joseph pareti]

Are you folks HP employees? if you are please get in touch with me joseph.pareti@hp.com