Another story of Oracle Support frustration

[gopi]

### Problem Description and its duration: ###
dbsnmp process is consuming gigs of memory and swap and eventually froze the database after filling the files system. The swap was 4 gig at it's highest. A trace file in network/trace was continually being written to and is 10gig by the time i stopped the agent. when i stopped the agent (agentctl stop agent) all the memory was release and the database freed.


### Oracle support response:
Database and application (query) tuning is an interactive process that requires a complete understanding of the environment where the database resides (database, operating system, application, etc). Standard product support services provides solutions to bugs that may affect the performance of the database and its components. Support may also provide general recommendations to start the tuning process.

racle support!

[DaPi]

Do we get an Oracle helicopter joke now?

(http://users.owt.com/lazrchet/humor/msjoke2.htm)

[slimdave]

Sounds like they're branching out into emotional support. Ask for the duty manager, 'cos this response seems wholey inappropriate

[stmontgo]

of course you should have received the standard
a) shutup reboot
b) shutup reinstall

which is typical from the EM group

[DaPi]

. . . and we pay for this "support" - unfortunately up-front.

[jodie]

gopi - did you ever get an answer to your problem?

We had a somewhat similar problem on AIX - although, now that I think about it, it was CPU, not Memory. Had to kill the unix snmp processes before we started the Oracle Agent. Apparently they don't get along well running at the same time.

HTH - Jodie

[gopi]

Sorry SlimDave for taking the wind out of your frustration thread but mine has reached a new level.

Next Oracle support told me to run a HANGANALYZE. I told them, the database is not hung anymore because I killed the processes and the swap freed up and the database runs fine. I also told them that when the agent is started, a trace file in network/trace is continually being written to. Meanwhile, the ironic part is that OMS is communicating with the agent just fine until memory and swap gets hosed.

Not knowing about HANGANALYZE, i asked does this thing work when the dB is not hung (I killed the I-agent). Oracle support answered:


Hi Frank,
No HANGANALYZE doesn't work when you are not seeing the problem. I thought you were still seeing frequent hanging.

Kindest Regards,
Tiffany


c.s. never addressed why the agent process chews up memory (leak) and the status of the tar is "Waiting on Customer".

Holy Sheee-it!!!

jodie...In order to really stop the agent, I too have to kill the unix processeses (2 of 'em).

[slimdave]

Originally posted by gopi
Sorry SlimDave for taking the wind out of your frustration thread but mine has reached a new level.

Don't worry, there's plenty of wind to go round.

Escalate to duty manager -- that's what got things moving on my own issue

[Mr.Hanky]

Ok, I confess!!

I really work for Oracle and thanks to you guys me and my associates have been laughing our arses off for days.

Try the Flux Capacitor



MH

[slimdave]

Originally posted by Mr.Hanky
Try the Flux Capacitor

Damn, he does work for Oracle Support!