On 04/24/2015 09:26 AM, Dominik Korittki wrote:
Hello all,
I am running two ipa3.3.3 instances in a replication on Centos 7 servers.
Last day the rootpartition went full (where the dirsrv databases are
stored), because of a big changelog-db.
dirsrv managed to do a graceful shutdown. Luckily, the second master
was still working properly, so i could recover the first one from it.
I resized the partition, booted up again and ran
'ipa-replica-manage re-initialize --from ipa02.internal'
Everything seemed to ran fine except for one warnig regarding an issue
with the changelog db, heres the log portion of the log
/var/log/dirsrv/slapd-INTERNAL/errors on recieving (first) IPA master:
[...]
[23/Apr/2015:10:41:46 +0200] NSMMReplicationPlugin -
multimaster_be_state_change: replica dc=internal is going offline;
disabling replication
[23/Apr/2015:10:41:47 +0200] - WARNING: Import is running with
nsslapd-db-private-import-mem on; No other process is allowed to
access the database
[23/Apr/2015:10:41:55 +0200] - import userRoot: Workers finished;
cleaning up...
[23/Apr/2015:10:41:55 +0200] - import userRoot: Workers cleaned up.
[23/Apr/2015:10:41:55 +0200] - import userRoot: Indexing complete.
Post-processing...
[23/Apr/2015:10:41:55 +0200] - import userRoot: Generating
numSubordinates complete.
[23/Apr/2015:10:41:55 +0200] - import userRoot: Flushing caches...
[23/Apr/2015:10:41:55 +0200] - import userRoot: Closing files...
[23/Apr/2015:10:41:55 +0200] - import userRoot: Import complete.
Processed 9983 entries in 8 seconds. (1247.88 entries/sec)
[23/Apr/2015:10:41:55 +0200] NSMMReplicationPlugin -
multimaster_be_state_change: replica dc=internal is coming online;
enabling replication
[23/Apr/2015:10:41:55 +0200] NSMMReplicationPlugin -
replica_reload_ruv: Warning: new data for replica dc=internal does not
match the data in the changelog.
Recreating the changelog file. This could affect replication with
replica's consumers in which case the consumers should be reinitialized.
[...]
this shouold be normal. at the moment of initialization, a server has a
database and a changelog. The datavase is recreated by initialization
and when replication plugin starts it detects that changelog and db no
longer match and recreates the changelog.
I am no expert in LDAP or Directory Server, but i noticed a
significant size difference of files in
/var/lib/dirsrv/slapd-INTERNAL/cldb/:
root@ipa01:~ > du -sch /var/lib/dirsrv/slapd-INTERNAL/cldb/*
0
/var/lib/dirsrv/slapd-INTERNAL/cldb/61e65983-718611e4-8059dc1c-48160578.sema
24M
/var/lib/dirsrv/slapd-INTERNAL/cldb/61e65983-718611e4-8059dc1c-48160578_546f4515000000040000.db
0
/var/lib/dirsrv/slapd-INTERNAL/cldb/9b310907-74a711e4-8059dc1c-48160578.sema
6,8M
/var/lib/dirsrv/slapd-INTERNAL/cldb/9b310907-74a711e4-8059dc1c-48160578_54748540000000600000.db
4,0K /var/lib/dirsrv/slapd-INTERNAL/cldb/DBVERSION
30M total
root@ipa02:/var/log > du -sch /var/lib/dirsrv/slapd-INTERNAL/cldb/*
0
/var/lib/dirsrv/slapd-INTERNAL/cldb/98ceaf89-74a711e4-910b9512-1512b1dc.sema
4,7G
/var/lib/dirsrv/slapd-INTERNAL/cldb/98ceaf89-74a711e4-910b9512-1512b1dc_546f4515000000040000.db
0
/var/lib/dirsrv/slapd-INTERNAL/cldb/9cfacd4b-74a711e4-910b9512-1512b1dc.sema
3,7M
/var/lib/dirsrv/slapd-INTERNAL/cldb/9cfacd4b-74a711e4-910b9512-1512b1dc_54748540000000600000.db
4,0K /var/lib/dirsrv/slapd-INTERNAL/cldb/DBVERSION
4,7G total
Also, i noticed a difference in the actual database size on both servers:
root@ipa01:~ > du -sch /var/lib/dirsrv/slapd-INTERNAL/db/*
4,0K /var/lib/dirsrv/slapd-INTERNAL/db/DBVERSION
1,3M /var/lib/dirsrv/slapd-INTERNAL/db/__db.001
544K /var/lib/dirsrv/slapd-INTERNAL/db/__db.002
9,6M /var/lib/dirsrv/slapd-INTERNAL/db/__db.003
1,4M /var/lib/dirsrv/slapd-INTERNAL/db/ipaca
2,2M /var/lib/dirsrv/slapd-INTERNAL/db/log.0000124384
101M /var/lib/dirsrv/slapd-INTERNAL/db/userRoot
115M total
root@ipa02:/var/log > du -sch /var/lib/dirsrv/slapd-INTERNAL/db/*
4,0K /var/lib/dirsrv/slapd-INTERNAL/db/DBVERSION
1,7M /var/lib/dirsrv/slapd-INTERNAL/db/__db.001
544K /var/lib/dirsrv/slapd-INTERNAL/db/__db.002
9,6M /var/lib/dirsrv/slapd-INTERNAL/db/__db.003
1,3M /var/lib/dirsrv/slapd-INTERNAL/db/ipaca
4,3M /var/lib/dirsrv/slapd-INTERNAL/db/log.0000074356
175M /var/lib/dirsrv/slapd-INTERNAL/db/userRoot
193M total
Besides from that, everything seems to be working fine again,
including the replication. No errors or warnings regarding this issue
are stated in dirsrv-logs. So I'm a bit confused right know wether to
believe everything worked fine or not.
Is this behaviour of IPA/Directory Server normal? Many thanks in advance!
Greetings and a nice day,
Dominik Korittki
--
Manage your subscription for the Freeipa-users mailing list:
https://www.redhat.com/mailman/listinfo/freeipa-users
Go to http://freeipa.org for more info on the project
