Hi,
since we updated to the latest CentOS 7 Version
# rpm -qa |
grep 389 389-ds-base-libs-1.3.11.1-3.el7_9.x86_64
389-ds-base-devel-1.3.11.1-3.el7_9.x86_64
389-ds-base-snmp-1.3.11.1-3.el7_9.x86_64
389-adminutil-devel-1.1.22-2.el7.x86_64
389-adminutil-1.1.22-2.el7.x86_64
389-ds-base-1.3.11.1-3.el7_9.x86_64
389-admin-1.1.46-4.el7.x86_64
3.10.0-1160.11.1.el7.x86_64
We experience strange locking (?) behaviour: we have a synchronisation jobs that tried to delete about 1300 Accounts, always 10 in parallel using some simple forking perl / shell scripts. Pstree locks like this
So we run 10 ldapmodify / ldapdelete calls nearly at the same time and the server does not do anything. After 100s to 400s it returns an error:
causing the client
Some modifies did work, but very slow either. Only a kill and
restart the ns-slapd helped. It's not strictly reproducable,
happens after a while...
Since we have other problems (see "389 DS memory growth") with
this version as well (the versions before did work perfectly for
years!) I think about upgrading the whole cluster to a debian
based system with a more recent Version on debian. We run also
some debian 11 based 389 and some IPAs in podman with Rocky and
have no problems at all.
Or are there any other hints we may could try to come around this strange behaviour on 1.3.11.1-3 ?
br
Harald