[Spacewalk-list] 500 ISE (Timeout) When Updating Kickstart Profile?

[Jan Pazdziora]

On Fri, Feb 17, 2012 at 02:06:56PM +0000, Wojtak, Greg (Superfly) wrote:
> Hi there,
> 
> Spacewalk 1.6
> Postgres
> Cent 6.2 (64-bit)
> 
> Has anyone else run into an issue where, when updating a kickstart profile (or cloning, or doing anything with it really) that you consistently time out with a 500 error after the request times out?
> 
> What I'm seeing looks to be nothing more than the query (queries?) taking too long to run.  I can see the queries while I'm waiting for a response after I hit the "Update Profile" button.  Eventually, when the request times out and the query finally finishes and I browse back into Spacewalk, I get the message at the top of the screen that the profile has been updated successfully (and indeed it appears it has been).
> 
> I've increased the HTTP timeout to 300 seconds (from 120) to try and work around this, but should it really take five minutes to update a profile?
> 
> The one query I was able to capture during this was:
> SELECT CP.package_id, CP.name_id, CP.evr_id, CP.package_arch_id FROM rhnPackageName PN inner join rhnChannelNewestPackage CP on CP.name_id = PN.id inner join rhnChannel C on C.id = Cp.channel_id  inner join rhnPackage P on P.id = CP.package_id inner join rhnPackageEvr EVR on P.evr_id = EVR.id  WHERE ( C.id = $1 or C.parent_channel = $2) AND PN.name = $3 AND C.label not like '%beta%' order by EVR.evr DESC
> 
> (sorry for the bad formatting)
> 
> I'm not sure what the $1, $2, and $3 are (I presume this is calling some sort of a procedure and those are arguments).
> 

Those are bind parameters. You can replace them with some reasonable
values.

Please analyze your tables and run explain plan to see what is
happening.

I assume your cobbler is running and is reachable just fine.

Anything of interest in /var/log/tomcat6/catalina.out?

-- 
Jan Pazdziora
Principal Software Engineer, Satellite Engineering, Red Hat