I've been asked to look at the performance of RHEL 7.1/RHCS 1.3. I keep running into these errors on 1 of my RHEL 7.1 client systems. The rbd devices are still present, but ceph-rbdname Is not in /usr/bin, but it is in trusty /usr/bin. Much like the rbdmap init script that ships with RHEL 7.1, but depends on functions from trusty /lib/lsb/init-functions (create a user defined systemd init function to map the images if rbd devices required at boot time) is this another example of RHEL 7.1 being not quite ready from prime time as a Ceph client? Can I ignore these messages? Or should I just return to my trusty client of choice and advise that to others? I'm going to want to know if these ceph-rbdnamer error paths are adding overhead to my performance testing on RHEL 7.1 clients so I will most like re-run everything with trusty clients to see for myself, but I'm curious what others have seen with RHEL/Centos/Fedora systemd Ceph clients.
There are 3 10TB rbd's in the cluster and 3 clients. Thanks. 14:22:43.018 Message from slave hd2_client0-0: 14:22:43.018 New messages found on /var/adm/messages. Do they belong to you? 14:22:43.018 /var/log/messages: Aug 5 15:22:39 essperf8 systemd-udevd: failed to execute '/usr/bin/ceph-rbdnamer' '/usr/bin/ceph-rbdnamer rbd2': No such file or directory 14:22:43.018 /var/log/messages: Aug 5 15:22:39 essperf8 systemd-udevd: failed to execute '/usr/bin/ceph-rbdnamer' '/usr/bin/ceph-rbdnamer rbd1': No such file or directory 14:22:43.018 /var/log/messages: Aug 5 15:22:39 essperf8 systemd-udevd: failed to execute '/usr/bin/ceph-rbdnamer' '/usr/bin/ceph-rbdnamer rbd1': No such file or directory
_______________________________________________ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
