[Freeipa-users] dead in the water IPA server

JR Aquino JR.Aquino at citrix.com
Tue May 15 05:36:01 UTC 2012


On May 13, 2012, at 2:39 PM, "Steven Jones" <Steven.Jones at vuw.ac.nz<mailto:Steven.Jones at vuw.ac.nz>> wrote:

Hi,

I have what I'm told are 6.3 rpms on ipa2 and no its not fixed, the memory leak kills a server in 48 hours.  I also find I have a problem with rebooting, IPA doesnt survive a reboot, so I cant even cron a reboot nightly.

Right now both are in a bad way and I need to reboot them......

:(

The interesting thing is I have a test setup that is stable, yet has the same rpms....so Im flumixt'd, maybe its something Ive done, but I cant think what....its bod standard as far as I know....

regards

Steven Jones

Technical Specialist - Linux RHCE

Victoria University, Wellington, NZ

0064 4 463 6272


When I was having similar problems, it turned out to be due to a few different factors...

* my cache was too low, was being exceeded and triggering a leak in 389
* I discovered a bug in managed entries that caused the plugin to fire if _any_ change occurred to a managed object. As opposed to firing only when relivent attributes changed.
* I also had a great deal of churning happening from slapi-nis in competition with the MemberOf plugin...

Here is my bug, it was fixed in Fedora, but perhaps it is still a problem in RHEL: https://bugzilla.redhat.com/show_bug.cgi?format=multiple&id=771493


________________________________
From: freeipa-users-bounces at redhat.com<mailto:freeipa-users-bounces at redhat.com> [freeipa-users-bounces at redhat.com<mailto:freeipa-users-bounces at redhat.com>] on behalf of Dmitri Pal [dpal at redhat.com<mailto:dpal at redhat.com>]
Sent: Saturday, 12 May 2012 9:29 a.m.
To: freeipa-users at redhat.com<mailto:freeipa-users at redhat.com>
Subject: Re: [Freeipa-users] dead in the water IPA server

On 05/07/2012 05:05 PM, Rich Megginson wrote:
On 05/07/2012 02:55 PM, Steven Jones wrote:

Hi,

Yes I have a memory leak see attached graphs....

Yes looks like the killer killed slapd.......dont know what caused this yet........if its the "killer" looks like its decided to kill slapd or slapd was going to kill the system anyway so it may have done the right thing.

Looks like I have 3 days between reboots if i dont IPA losses the plot big time....very bad news..........I will I think slow IPA deployment here at this time........this cant be deployed for us as it is, I cant even test as if something doesn't work I don't know if its my configuring error or an inconsistent IPA.

:/

Thanks for this info I will pursue this through RH support for a perm fix, adding more memory doesn't strike me as the solution, 4gb of ram for 3~4 users and about 6 client machines seems a lot.....

Right.  See https://fedorahosted.org/389/ticket/51 and especially all of the comments to https://bugzilla.redhat.com/show_bug.cgi?id=697701

You will need to closely monitor your entry cache usage.


As far as I see the ticket is fixed upstream and is in testing for 6.3.
Is this the correct understanding?


regards

Steven Jones

Technical Specialist - Linux RHCE

Victoria University, Wellington, NZ

0064 4 463 6272

________________________________________
From: Sigbjorn Lie [sigbjorn at nixtra.com<mailto:sigbjorn at nixtra.com>]
Sent: Monday, 7 May 2012 9:45 p.m.
To: Steven Jones
Cc: Jan Cholasta; freeipa-users at redhat.com<mailto:freeipa-users at redhat.com>
Subject: Re: [Freeipa-users] dead in the water IPA server

This sound very much the same as the issue I've been having. Did you check to see if it was the
directory server that consumed all of your memory too?

https://www.redhat.com/archives/freeipa-users/2012-April/msg00139.html


Regards,
Siggi




On Mon, May 7, 2012 11:32, Jan Cholasta wrote:


Hi,


It seems that your system ate all the available memory and the kernel
decided to kill a directory server instance to free some. The kernel agent responsible for this is
called the out-of-memory killer, you can read more about it and how to configure it not to kill
important processes here: http://lwn.net/Articles/317814/

On 7.5.2012 02:22, Steven Jones wrote:



Interesting memory message.....as attached....


I take it it isnt good?  cant login that is for sure so whatever is behind the web gui is dead
if nothing else...


regards

Steven Jones


Technical Specialist - Linux RHCE


Victoria University, Wellington, NZ


0064 4 463 6272




--
Jan Cholasta


_______________________________________________
Freeipa-users mailing list
Freeipa-users at redhat.com<mailto:Freeipa-users at redhat.com>
https://www.redhat.com/mailman/listinfo/freeipa-users






_______________________________________________
Freeipa-users mailing list
Freeipa-users at redhat.com<mailto:Freeipa-users at redhat.com>
https://www.redhat.com/mailman/listinfo/freeipa-users



_______________________________________________
Freeipa-users mailing list
Freeipa-users at redhat.com<mailto:Freeipa-users at redhat.com>
https://www.redhat.com/mailman/listinfo/freeipa-users



--
Thank you,
Dmitri Pal

Sr. Engineering Manager IPA project,
Red Hat Inc.


-------------------------------
Looking to carve out IT costs?
www.redhat.com/carveoutcosts/<http://www.redhat.com/carveoutcosts/>




<ipa2-memory-error-12.jpeg>
<ipa1-memory-error-14.jpeg>
<ipa2-cpu-error-11.jpeg>
_______________________________________________
Freeipa-users mailing list
Freeipa-users at redhat.com<mailto:Freeipa-users at redhat.com>
https://www.redhat.com/mailman/listinfo/freeipa-users




More information about the Freeipa-users mailing list