Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
/ 2006-09-28 18:02:22 +0200
\ KOZMAN Balint:
>
> Hi All,
>
>
> I have a two-primaries setup with three devices (drbd{0,1,2}). After an
> unexpected shutdown of node2, the following situation occured:
>
> resource provided by drbd0 on node1 is:
> 0: cs:StandAlone st:Primary/Unknown ds:UpToDate/DUnknown r---
> ns:0 nr:0 dw:14484 dr:1375443 al:0 bm:348 lo:0 pe:0 ua:0 ap:0
> resync: used:0/7 hits:65701 misses:91 starving:0 dirty:0 changed:91
> act_log: used:0/257 hits:3590 misses:0 starving:0 dirty:0 changed:0
>
> on node2 it's:
> 0: cs:StandAlone st:Secondary/Unknown ds:Inconsistent/DUnknown r---
> ns:0 nr:0 dw:0 dr:0 al:0 bm:257 lo:0 pe:0 ua:0 ap:0
> resync: used:0/7 hits:0 misses:0 starving:0 dirty:0 changed:0
> act_log: used:0/257 hits:0 misses:0 starving:0 dirty:0 changed:0
>
>
> I would like to make node2 resync its data from node1, but after
> connecting them, they fall back to standalone mode. I tried this on node2:
>
> drbdsetup /dev/drbd0 net 10.0.0.11:7788 10.0.0.10:7788 C -D
>
> but the problem remains after reinitiating the connection. Also strange is
> that the kernel log of node1 writes:
>
> drbd0: State change failed: Refusing to be Primary without at least one UpToDate disk
> drbd0: state = { cs:StandAlone st:Primary/Unknown ds:UpToDate/DUnknown r--- }
> drbd0: wanted = { cs:StandAlone st:Primary/Unknown ds:Outdated/DUnknown r--- }
> drbd0: outdate-peer helper broken, returned 0
> drbd0: receiver terminated
>
> which I just cannot understand. As node1 _is_ primary and uptodate
> according to the output of /proc/drbd.
node2 tries to become primary. to do so, it first has to "resource
level fence" node1. so it tries to "outdate" node1. and node1
correctly refuses to be outdated.
the error message should better be written as
"Refusing to throw away the last good accesible data copy while being Primary"
but, seemingly, your outdate-peer-helper is broken.
> Any ideas?
try with current svn. maybe thats better. maybe it won't work at all.
if you look at the commit log, you'll notice we have been sort of busy
working on drbd8 for the last few weeks. we expect to have pre5 early
next week, and a rc1 mid-october.
though conflict detection for two-primary and resync on reconnection of
two-primaries is not yet correct for all corner cases. that will be
sorted out later.
--
: Lars Ellenberg Tel +43-1-8178292-0 :
: LINBIT Information Technologies GmbH Fax +43-1-8178292-82 :
: Schoenbrunner Str. 244, A-1120 Vienna/Europe http://www.linbit.com :
__
please use the "List-Reply" function of your email client.