Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Hi Lars,
Thanks for the suggestion to upgrade from 8.0.2 to 8.0.3.
Unfortunately this hasn't fixed my problem. Writing to partitions
whilst in re-sync still results in DRBD becoming 'stalled'. Which, as I
explained before, is a terminal condition requiring a power-cycle to
unlock. However this does not occur with protocol 'C':
version: 8.0.3 (api:86/proto:86)
SVN Revision: 2881 build by root at hp-tm-02, 2007-05-15 10:38:31
0: cs:SyncSource st:Primary/Secondary ds:UpToDate/Inconsistent B r---
ns:423428 nr:140 dw:95888 dr:334423 al:27 bm:141 lo:2 pe:0 ua:0 ap:1
[>...................] sync'ed: 3.2% (9924/10244)M
stalled
resync: used:1/31 hits:20493 misses:34 starving:0 dirty:0
changed:34
act_log: used:3/257 hits:23910 misses:42 starving:0 dirty:15
changed:27
1: cs:SyncTarget st:Secondary/Primary ds:Inconsistent/UpToDate B r---
ns:876 nr:923328 dw:924204 dr:4529 al:0 bm:99 lo:0 pe:0 ua:0 ap:0
[>...................] sync'ed: 0.4% (10212/10244)M
stalled
resync: used:1/31 hits:3855 misses:17 starving:0 dirty:0 changed:17
act_log: used:1/257 hits:219 misses:0 starving:0 dirty:0 changed:0
Sorry to report such errors!
Just another very small error. The 'man' for drbdadm reports the
command: 'invalidate_remote'. This should of course be 'invalidate-remote'.
Thanks for the help, I hope it's not too much of a problem,
Ben
Lars Ellenberg wrote:
> On Tue, May 15, 2007 at 08:11:27AM +0100, Ben Clewett wrote:
>> DRBD,
>>
>> After my discovery of the 'stalled' problem I later found a solution:
>>
>> I changed my protocol from 'B' to 'C' and the problem went away.
>>
>> I would therefore suggest that there is a bug in the 'B' protocol in
>> version 8.0.2.
>>
>
> there is a bug in 8.0.2 which is more likely to trigger with protocol B.
> thats why there is 8.0.3.
>