Avoiding the merits of SysV vs SystemD: I went and grabbed the systemd init scripts (unit files, whatever) from upstream. As Christian suggests, answers will vary depending on your ceph version (and where your packages came from), but if you go to: https://github.com/ceph/ceph/tree/master/systemd and grab *{service,target} and place those in the systemd directory (/usr/lib/systemd/system/ on my EL systems), everything works *a lot* better. We're still on Hammer, so the above scripts required some minor modifications (basically, they added a ceph user account, but Hammer doesn't know about it, so you have to remove the parts of the command-line that specify user accounts). Bill On Mon, Mar 7, 2016 at 3:44 AM, Florent B <florent@xxxxxxxxxxx> wrote: > Hi, > > Yes sorry version are Jessie and last Infernalis package 9.2.1-1~bpo80+1. > > I could switch to SysV, but I don't want to, only for Ceph. All my > others services are working well, I don't know why Ceph couldn't. > > On 03/07/2016 12:41 PM, Christian Balzer wrote: >> Hello, >> >> since everybody else who might actually have answers for you will ask >> this: >> >> What version of Debian (Jessie one assumes)? >> What Ceph packages, Debian ones or from the Ceph repository? >> Exact versions please. >> >> As for me, I had similar experiences with Firefly (Debian package) under >> Jessie and switched to SysV init for serene happiness. >> >> Christian >> >> On Mon, 7 Mar 2016 12:27:48 +0100 Florent B wrote: >> >>> Hi everyone, >>> >>> I try to understand how Ceph works with systemctl on Debian and it seems >>> to be a mess. >>> >>> First question : why is /etc/init.d/ceph is used as a "ceph" service ? >>> that's an old sysvinit script ! >>> >>> Second question: why this "ceph" service is handling everything while it >>> should be the new "ceph.target" unit ? "ceph.target" is disabled ! >>> >>> Third question : why some components like OSDs are handled via systemctl >>> whereas MONs are unable to be handled via systemctl ? Here examples : >>> >>> # systemctl status ceph-mon@* -l >>> ● ceph-mon@\x2a.service - Ceph cluster monitor daemon >>> Loaded: loaded (/lib/systemd/system/ceph-mon@.service; enabled) >>> *Active: failed (Result: exit-code) *since Mon 2016-03-07 12:16:50 >>> CET; 1min 55s ago >>> Process: 1861 ExecStart=/usr/bin/ceph-mon -f --cluster ${CLUSTER} --id >>> %i --setuser ceph --setgroup ceph (code=exited, status=1/FAILURE) >>> Main PID: 1861 (code=exited, status=1/FAILURE) >>> >>> ● ceph-mon@5.service - Ceph cluster monitor daemon >>> Loaded: loaded (/lib/systemd/system/ceph-mon@.service; enabled) >>> *Active: failed (Result: exit-code)* since Tue 2016-03-01 12:46:27 >>> CET; 5 days ago >>> Main PID: 10600 (code=exited, status=1/FAILURE) >>> >>> ● ceph-mon@host2.service - Ceph cluster monitor daemon >>> Loaded: loaded (/lib/systemd/system/ceph-mon@.service; enabled) >>> *Active: failed (Result: exit-code)* since Mon 2016-03-07 12:07:48 >>> CET; 10min ago >>> Process: 31231 ExecStart=/usr/bin/ceph-mon -f --cluster ${CLUSTER} >>> --id %i --setuser ceph --setgroup ceph (code=exited, status=1/FAILURE) >>> Main PID: 31231 (code=exited, status=1/FAILURE) >>> >>> >>> # systemctl status ceph-osd@* >>> ● ceph-osd@15.service - Ceph object storage daemon >>> Loaded: loaded (/lib/systemd/system/ceph-osd@.service; disabled) >>> *Active: active (running)* since Fri 2016-02-26 19:04:45 CET; 1 weeks >>> 2 days ago >>> Main PID: 30211 (ceph-osd) >>> CGroup: /system.slice/system-ceph\x2dosd.slice/ceph-osd@15.service >>> └─30211 /usr/bin/ceph-osd -f --cluster ceph --id 15 --setuser >>> ceph... >>> >>> ● ceph-osd@14.service - Ceph object storage daemon >>> Loaded: loaded (/lib/systemd/system/ceph-osd@.service; disabled) >>> *Active: active (running)* since Tue 2016-03-01 12:44:32 CET; 5 days >>> ago Main PID: 9586 (ceph-osd) >>> CGroup: /system.slice/system-ceph\x2dosd.slice/ceph-osd@14.service >>> └─9586 /usr/bin/ceph-osd -f --cluster ceph --id 14 --setuser >>> ceph ... >>> >>> ceph-osd ID's are automatically found, but ceph-mon ID's are not ! >>> >>> "systemctl start ceph.target" is not starting MONs on my systems ! >>> >>> What's in trouble ? How to handle this mess ? >>> >>> Thank you for your help >>> >>> Florent >> > > _______________________________________________ > ceph-users mailing list > ceph-users@xxxxxxxxxxxxxx > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |