We had a problem today, one of our two 389 DS servers hanged showing the errors:
ERR - libdb - BDB2055 Lock table is out of available lock entries ERR - NSMMReplicationPlugin - changelog program - _cl5CompactDBs - Failed to compact db5f7bb9-ab0611e6-9bc8987f-40ec05bf; db error - 12 Cannot allocate memory
We redirected all the queries to our second 389 DS Server and started debugging the problem.
It's the second time that this happens to us. We did our homework the first time, four months ago, and didn't found any problem.
At that time, we just increased the number of open files and reviewed the documentation about tunning nsslapd-db-locks (https://access.redhat.com/solutions/3217591). We set the nsslapd-db-locks to 100000 as recommended.
But today, with a little more time, we reviewed the documentation about compacting the log manually (https://access.redhat.com/documentation/en-us/red_hat_directory_server/10/ht...).
We started to increase the nsslapd-db-locks, first up to 200000, then to half a million, and at the end to 1.5 millions.
Every check failed but the last one. We have no error logs (nor an affirmative one), but the database file in the 'changelogdb' folder went from 13GB to 2.8GB.
Is it reasonable to keep nsslapd-db-locks so high?
On 4/26/21 8:33 AM, murmansk@hotmail.com wrote:
We had a problem today, one of our two 389 DS servers hanged showing the errors:
ERR - libdb - BDB2055 Lock table is out of available lock entries ERR - NSMMReplicationPlugin - changelog program - _cl5CompactDBs - Failed to compact db5f7bb9-ab0611e6-9bc8987f-40ec05bf; db error - 12 Cannot allocate memory
We redirected all the queries to our second 389 DS Server and started debugging the problem.
It's the second time that this happens to us. We did our homework the first time, four months ago, and didn't found any problem.
At that time, we just increased the number of open files and reviewed the documentation about tunning nsslapd-db-locks (https://access.redhat.com/solutions/3217591). We set the nsslapd-db-locks to 100000 as recommended.
But today, with a little more time, we reviewed the documentation about compacting the log manually (https://access.redhat.com/documentation/en-us/red_hat_directory_server/10/ht...).
We started to increase the nsslapd-db-locks, first up to 200000, then to half a million, and at the end to 1.5 millions.
Every check failed but the last one. We have no error logs (nor an affirmative one), but the database file in the 'changelogdb' folder went from 13GB to 2.8GB.
Is it reasonable to keep nsslapd-db-locks so high?
There is a bit of debate on this topic, but really the only impact of more db locks is that one of the private DB files will be larger (but it's not significant in size). IMO I would say it's perfectly safe to keep it at that value.
Mark
389-users mailing list -- 389-users@lists.fedoraproject.org To unsubscribe send an email to 389-users-leave@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/389-users@lists.fedoraproject.... Do not reply to spam on the list, report it: https://pagure.io/fedora-infrastructure
Thanks for your answer Mark!!
We'll configure the same value on our second server.
I don’t know about increasing dblocks, but that seems like the same crash that was affecting our systems as well. We had two identical crashes like that, and I recently updated all the packages in hopes that it was just a bad linked library.
If we have the same crash after updating I’ll try and find more data. I could get the crash dump data last time.
On Mon, Apr 26, 2021 at 23:20 murmansk@hotmail.com wrote:
Thanks for your answer Mark!!
We'll configure the same value on our second server. _______________________________________________ 389-users mailing list -- 389-users@lists.fedoraproject.org To unsubscribe send an email to 389-users-leave@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/389-users@lists.fedoraproject.... Do not reply to spam on the list, report it: https://pagure.io/fedora-infrastructure
389-users@lists.fedoraproject.org