[DRBD-user] DRBD9 drbdadm complains about fencing being in wrong section

Igor Cicimov icicimov at gmail.com
Fri Feb 26 00:16:14 CET 2016

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


On Mon, Feb 22, 2016 at 9:36 PM, Igor Cicimov <icicimov at gmail.com> wrote:

> Sorry, forgot to include the list in my reply.
>
>
> ---------- Forwarded message ----------
> From: Igor Cicimov <icicimov at gmail.com>
> Date: Mon, Feb 22, 2016 at 9:33 PM
> Subject: Re: [DRBD-user] DRBD9 drbdadm complains about fencing being in
> wrong section
> To: Roland Kammerer <roland.kammerer at linbit.com>
>
>
> Hi Roland,
>
> On Thu, Feb 18, 2016 at 8:03 PM, Roland Kammerer <
> roland.kammerer at linbit.com> wrote:
>
>> On Thu, Feb 18, 2016 at 10:37:58AM +1100, Igor Cicimov wrote:
>> > On Wed, Feb 17, 2016 at 8:18 PM, Roland Kammerer <
>> roland.kammerer at linbit.com
>> > > wrote:
>> >
>> > > On Wed, Feb 17, 2016 at 04:20:12PM +1100, Igor Cicimov wrote:
>> > > > Hi,
>> > > >
>> > > > I'm testing 9.0.1.1 installed from git and have a resource with
>> fencing
>> > > in
>> > > > the disk section:
>> > > >
>> > > >     disk {
>> > > >         on-io-error detach;
>> > > >         fencing resource-and-stonith;
>> > > >     }
>> > >
>> > > It belongs to net{}, and yes, the man page is outdated, I will fix
>> that.
>> > >
>> > >
>> >  Thanks Roland that worked. By the way I'm facing another issue when
>> > starting the service:
>> >
>> > # service drbd start
>> >  * Starting DRBD
>> > resources
>> > [
>> >      create res: vg1
>> >    prepare disk: vg1
>> >     adjust disk: vg1
>> >     prepare net: vg1
>> > adjust peer_devices: vg1
>> > attempt to connect: vg1
>> > ]
>> > ...drbdadm: Unknown command 'sh-b-pri'
>>
>> Too old version of drbd-utils (which provides the init script)?
>>
>
> Nope, latest install from git as it can be seen in the output of the
> drbdadm version in my initial post.
>
> I'm also seeing another discrepancy between the utility and what is in the
> docs:
>
> # drbdadm status vg1 --verbose --statistics
> drbdadm: unrecognized option '--statistics'
> try 'drbdadm help'
>
> where '--statistics' is suppose to be a valid option according to the docs.
>
>
>> 'sh-b-pri' was dropped, see the corresponding comment in the init
>> script:
>> # Become primary if configured
>> # Currently, this is necessary for drbd8
>> # drbd9 supports automatic promote and removes the
>> # "sh-b-pri" command.
>> $DRBDADM sh-b-pri all || true
>>
>> >
>> > although drbdadm tells me that all is fine:
>> >
>> > # drbdadm status
>> > vg1 role:Primary
>> >   disk:UpToDate
>> >   drbd02 role:Primary
>> >     peer-disk:UpToDate
>>
>> Looks good.
>>
>> >
>> > Now this is a problem since I can't set drbd under pacemaker control:
>> >
>> > Failed actions:
>> >     p_drbd_vg1_start_0 (node=drbd01, call=9, rc=1, status=Timed Out,
>> > last-rc-change=Thu Feb 18 10:07:43 2016
>> > , queued=20004ms, exec=1ms
>> > ): unknown error
>> >     p_drbd_vg1_start_0 (node=drbd02, call=9, rc=1, status=Timed Out,
>> > last-rc-change=Thu Feb 18 10:07:43 2016
>> > , queued=20005ms, exec=0ms
>> > ): unknown error
>> >
>> > I've been running this kind of setups with 8.4.4 with no issues. Has
>> > something changed around drbd management in version 9 or is it just
>> > something wrong with the initd script?
>>
>> AFAIK you should not mix the drbd service with pacemaker anyways. Choose
>> one, pacemaker. Pacemaker is really not my field of expertise, so
>> somebody else has to jump in here.
>>
>>
> I'm not mixing them drbd is set to not autostart and should be under
> pacemaker control...eventually.
>
>
>> > Finally thanks for the drbdmanage pointer I'll do some research about
>> it (I
>> > think I saw some presentation of yours on the web). I don't see it
>> > installed on my system (Ubuntu 14.04.4 LTS) so guess it is not packaged
>> > with drbd-utils but is a separate toll. I have a question though, since
>> I
>> > have done the config already manually, is the drbdmanage going to take
>> over
>> > based on already existing configuration or I will need to reconfigure
>> > everything all over again?
>>
>> It is a separate package and not part of drbd-utils. AFAIR downstream
>> Debian/Ubuntu did not pick it up yet. For Ubuntu there is a PPA if you
>> want to try it:
>>
>> https://launchpad.net/~linbit/+archive/ubuntu/linbit-drbd9-stack
>>
>> Currently there is no importing of manually crafted res files. To me, it
>> is actually very low on the feature list I care about. Depending on the
>> size/time for resync/... it is probably the easiest if you start from
>> scratch, but then it is really easy:
>>
>> node1-n:$ vgcreate drbdpool ...
>> node1:$ drbdmanage init
>> node1:$ drbdmange add-node node2 IP
>> node1:$ drbdmange new-volume foo 10G --deploy 2
>>
>> http://drbd.linbit.com/users-guide-9.0/ch-admin-drbdmanage.html
>>
>> And you are done, lvm configured, DRBD on top of it.
>>
>> Regards, rck
>> _______________________________________________
>> drbd-user mailing list
>> drbd-user at lists.linbit.com
>> http://lists.linbit.com/mailman/listinfo/drbd-user
>>
>
>
>
And one more thing I find confusing. There is no drbd port listening for
connections on the drbd nodes in the netstats output which I find quite
strange. Maybe that is causing the pacemaker timeouts while trying to
connect to drbd.

Following is the relevant part of the config:

    volume 0 {
       device       /dev/drbd0;
       disk           /dev/sda;
       meta-disk   internal;
    }
    on drbd01 {
       address     10.1.1.152:7788;
    }
    on drbd02 {
       address     10.1.1.198:7788;
    }
    connection-mesh {
       hosts         drbd01 drbd02;
    }

Thanks,
Igor
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20160226/05d31b00/attachment.htm>


More information about the drbd-user mailing list