On 06/22/2015 10:22 AM, Tamas Papp wrote:

On 06/19/2015 11:12 AM, Christoph Kaminski wrote:

for this problem you can see the thread "Haunted servers?" here on ml. There is a solution from me for this but it doesnt work 100% :/

I would rather rerun the replication.

we have a Ticket @Red Hat for this problem, (https://access.redhat.com/support/cases/#/case/01429034if you have rh support)
But is really sad/silly how RH support works (read the whole ticket).

Unfortunately I don't have access there.

In fact we have a bigger issue here, but I don't know, if it's related.

The whole story is the following:

I migrated (ipa migrate-ds) about 150 users between two ldap databases. Old one was v3.0 (centos 6.6), the new one is v4.1 (centos 7.1). After migrating users I switched off old servers and replaced centos 6.6 machines with centos 7.1. Than replica servers was installed. One replicas had to be reinstalled one time, because the replica process was hanged up for some reason. Now two servers (not the reinstalled one, but the original master and one other) crash quite frequently with sigsegv. It's like something is leaking.

I tuned the the nsslapd-cachememsize value of the following config entries:

dn: cn=changelog,cn=ldbm database,cn=plugins,cn=config
dn: cn=config, cn=ldbm database, cn=plugins, cn=config
dn: cn=userRoot,cn=ldbm database,cn=plugins,cn=config

It helped a lot, but not enough.

Hello Tamas,

You get some sigsegv and you may hit a real bug. Please try to capture a core (http://www.port389.org/docs/389ds/FAQ/faq.html#debugging-crashes) then you may attach a pstack of it.


I had seen this before years ago, when I started using ipa. It was on Fedora and the bleeding edge Freeipa version. At that time I switched to CentOS because I trusted more in the well tested enterprise distribution.
But now it's not an option:)

Any suggestion?


Manage your subscription for the Freeipa-users mailing list:
Go to http://freeipa.org for more info on the project

Reply via email to