> -----Ursprüngliche Nachricht----- > Von: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] > ha.org] Im Auftrag von Serge Dubrouski > Gesendet: Donnerstag, 31. Juli 2008 18:55 > An: General Linux-HA mailing list > Betreff: Re: [Linux-HA] samba lsb script > > On Thu, Jul 31, 2008 at 10:48 AM, Serge Dubrouski <[EMAIL PROTECTED]> wrote: > > On Thu, Jul 31, 2008 at 10:43 AM, Schmidt, Florian > > <[EMAIL PROTECTED]> wrote: > >> > >> > >>> -----Ursprüngliche Nachricht----- > >>> Von: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] > >>> ha.org] Im Auftrag von Serge Dubrouski > >>> Gesendet: Donnerstag, 31. Juli 2008 16:21 > >>> An: General Linux-HA mailing list > >>> Betreff: Re: [Linux-HA] samba lsb script > >>> > >>> On Thu, Jul 31, 2008 at 7:24 AM, Schmidt, Florian > >>> <[EMAIL PROTECTED]> wrote: > >>> > Hi list, > >>> > > >>> > Samba is part of my linux-ha and drbd setup. > >>> > > >>> > Because I want to store the config files at a central place, I put them > >>> > onto a > >>> DRBD-device. > >>> > > >>> > Problem is: when heartbeat starts it first triggers a monitor operation > >>> > of all > >>> resources. > >>> > Because samba cannot find its config files (they are on the > >>> > drbd-device, that > is > >>> not mounted yet, with a symlink from /etc/samba to /configs/samba) it > >>> returns > 4. > >>> > >>> Is Samba a part of your HA cluster? If so why do you start your > >>> resources in that order? I'd try to make a group of IP, DRBD, > >>> Filesystem and Samba resources in which Samba would be the last one to > >>> start. > >> > >> Cause the setup is a little more complicated ;) Samba is only a minor > >> service, > that needs to run. > >> > >> I do have these constraints: > >> <rsc_colocation id="Run_samba_only_on_Group_configs_started" > from="samba" to="Group_configs" to_role="started" score="INFINITY"/> > >> <rsc_order id="Start_samba_after_Group_configs" action="start" > from="samba" type="after" to="Group_configs" to_action="start"/> > >> <rsc_order id="Start_samba_after_Group_ftpdata" action="start" > from="samba" type="after" to="Group_ftpdata" to_action="start"/> > > > > Ahhh! The problem is that Pacemaker needs to check that resource isn't > > already up before starting it. That's where it fails. So I think that > > your only option is a custom samba script that will return a proper > > exit code (resource is down) if config directory isn't mounted.
Exactely. :) Well, uncommenting made it work and I got no further errors until now :) > Could you post your log files? Still I don't get why it tries to > start/monitor samba before Group_configs and Group_ftpdata are up. It does not try to start samba before they are up. It only wants to do one monitor-action to be sure it's not started yet. I think normally the init-script status action should not check if config-files are available but only check if there's a pidfile and smbd/nmbd-process etc. My log file is attached to the first mail. Around line 500 it's interesting :) Have a nice evening. Florian > > > > >> > >> which, in my opinion, means: wait until both Filesystems are mounted (each > group contains on drbddisk and one Filesystem-RA) and start on that node, > where > the configs-filesystem is mounted, because your config-files are located > there. > >> > >> But this doesn't avoid heartbeat from triggering a samba-monitoring action > >> at > startup, which should return, that the resource isn't running yet. Because > the smb- > init-script looks for its config-files and doesn't find them yet, it fails > and return a > wrong code. > >> > >> I simply commented out the 2 lines, where it checks if its config-file is > >> at its place > and now it is running as expected. > >> > >> > >> Thanks for your advice > >> Florian > >> > >> > >>> >So heartbeat does not know how to handle this and does not start the > resource. > >>> > > >>> > Does anyone have a script that handles this fine or do I have to make > >>> > the > init- > >>> script more lsb-compatible? > >>> > > >>> > > >>> > Regards > >>> > Florian > >>> > > >>> > PS: The startup log and /etc/init.d/smb are attached. > >>> > > >>> > _______________________________________________ > >>> > Linux-HA mailing list > >>> > [email protected] > >>> > http://lists.linux-ha.org/mailman/listinfo/linux-ha > >>> > See also: http://linux-ha.org/ReportingProblems > >>> > > >>> > >>> > >>> > >>> -- > >>> Serge Dubrouski. > >>> _______________________________________________ > >>> Linux-HA mailing list > >>> [email protected] > >>> http://lists.linux-ha.org/mailman/listinfo/linux-ha > >>> See also: http://linux-ha.org/ReportingProblems > >> _______________________________________________ > >> Linux-HA mailing list > >> [email protected] > >> http://lists.linux-ha.org/mailman/listinfo/linux-ha > >> See also: http://linux-ha.org/ReportingProblems > >> > > > > > > > > -- > > Serge Dubrouski. > > > > > > -- > Serge Dubrouski. > _______________________________________________ > Linux-HA mailing list > [email protected] > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
