Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
please subscribe. currently I have to explicitly let your messages
through by hand. costs time. please don't send html.
thanks.
> > > I have installed DRBD-0.7.10 on AS3, ha1 & ha2.
> >
> > you want to use 0.7.13, in case you have SMP.
> >
> > > ha1 is the primary and there are many files in it.
> > > ha2 is the secondary, it will be replicated after I start up DRBD.
> > >
> > > I have checked that ha1 is primary and ha2 is secondary before I
> > > connected the cross link cable between them. All files in ha1 are
> > > gone after I connected this cable. I don't know what is going on.
> > > This isn't the first time I installed DRBD. I have done before and
> > > they were OK then.
> > >
> > > Which commands I can use to force data sync from ha1 to ha2 while these
> > > two ha1 & ha2 are connected firstly, and that can prevent me from
> > > this mistake?
> >
> > what do you mean by "gone"? where exactly have "all files" been before?
>
>
> DRBD is a very good software for HA, and replicated data automatically.
> I just recover host and DRBD will take care of HA, and I don't worry
> anything.
>
> To initialize DRBD first time, it will automatically sync data from primary
> to
> secondary. But sometimes it sync data from secondary to primary,
> if I made some procedures wrong while starting DRBD.
drbd will _never_ sync to a node that is in primary state.
it is possible to make a synctarget primary.
but it is not possible to make a primary sync target.
don't confuse what you call the boxes, and which state they are in.
> I have installed DRBD for many times, and it works well. I knew I did
> something wrong this time. Does DRBD provide some commands to sync
> data from primary to secondary manually? and make sure data to be
> replicated first to prevent mistakes done by person.
why not read the manpage?
drbdadm invalidate
> ha1: /dev/sdb1 : there are many files in it before started DRBD.
> ha2: /dev/sda4 : did 'mkfs -t ext3 /dev/sda4' before started DRBD.
I really hope you use external meta data.
even then, you should not use the lower level devices.
so, if you want to create a file system, you "mkfs /dev/drbd0"
(obviously this is only possible when Primary).
on the other node (the gonna-be sync-target), all effects of anything
are overwritten by the synchronization. so it is nonsense.
> I started both DRBD, and the DRBD sync data from ha2 /dev/sda4 to
> ha1 /dev/sdb1, and clean /dev/sda4 of ha1. I want to sync from ha1 to ha2.
>
> > where is the drbd mounted?
>
>
> on ha1 {
> device /dev/drbd0;
> disk /dev/sdb1;
> address 192.168.0.192:7788;
> meta-disk internal;
> }
this has nothing to do with the question (mount point?).
but it is interessting, nonetheless.
to quote from the example drbd.conf:
# NOTE that if you do not have some dedicated partition to use for
# the meta-data, you may use 'internal' meta-data.
#
# THIS HOWEVER WILL DESTROY THE LAST 128M
# OF THE LOWER LEVEL DEVICE.
#
# So you better make sure you shrink the filesystem by 128M FIRST!
# or by 132M just to be sure... :)
do you notice something?
> > do you have heartbeat configured?
> > did your actions interfere with heartbeat actions?
>
>
> I have heartbeat then.
> ---------------------- haresources
> --------------------------------------------
> ha1 drbddisk::r0 Filesystem::/dev/drbd0::/raid::ext3
> 210.59.22.14
> haj
so you told heartbeat to mount /dev/drbd0 on /raid
btw, you probably should list the ip first on the resource line.
> > please provide some log messages from that event.
>
> Sep 10 14:19:56 ha1 kernel: drbd: initialised. Version: 0.7.10(api:77/proto:74)
> Sep 10 14:19:56 ha1 kernel: drbd: SVN Revision: 1743 build by root at baumeister, 2005-01-31 13:08:05
> Sep 10 14:19:56 ha1 kernel: drbd: registered as block device major 147
> Sep 10 14:19:56 ha1 kernel: drbd0: Creating state block
> drbd0: resync bitmap: bits=137184413 words=4287014
> drbd0: size = 523 GB (548737652 KB)
> drbd0: Assuming that all blocks are out of sync (aka FullSync)
> drbd0: 548737652 KB now marked out-of-sync by on disk bit-map.
> drbd0: drbdsetup [2809]: cstate Unconfigured --> StandAlone
> drbd0: drbdsetup [2844]: cstate StandAlone --> Unconnected
> drbd0: drbd0_receiver [2845]: cstate Unconnected --> WFConnection
> drbd0: Secondary/Unknown --> Primary/Unknown
> drbd0: Primary/Unknown --> Secondary/Unknown
> drbd0: Secondary/Unknown --> Primary/Unknown
>
> Sep 10 14:24:58
> ha1 kernel: drbd0: Handshake successful: DRBD Network Protocol version 74
> drbd0: Connection established.
> drbd0: I am(P): 1:00000002:00000001:00000001:00000002:10
> drbd0: Peer(S): 1:00000006:00000004:00000027:00000006:01
> drbd0: Current Primary shall become sync TARGET! Aborting to prevent data corruption.
ok. I guess you understand plain english.
> drbd0: drbd0_receiver [2845]: cstate WFReportParams --> StandAlone
> drbd0: error receiving ReportParams, l: 72!
> drbd0: worker terminated
> drbd0: asender terminated
> drbd0: drbd0_receiver [2845]: cstate StandAlone --> StandAlone
> drbd0: Connection lost.
> drbd0: receiver terminated
next try.
> Sep 10 14:32:50 ha1 drbd: Starting DRBD resources:
> Sep 10 14:32:50 ha1 kernel: drbd: initialised. Version: 0.7.10(api:77/proto:74)
> Sep 10 14:32:50 ha1 kernel: drbd: SVN Revision: 1743 build by root at baumeister, 2005-01-31 13:08:05
> Sep 10 14:32:50 ha1 kernel: drbd: registered as block device major 147
> Sep 10 14:32:50 ha1 kernel: drbd0: resync bitmap: bits=137184413 words=4287014
> drbd0: size = 523 GB (548737652 KB)
> drbd0: 523 GB marked out-of-sync by on disk bit-map.
> drbd0: Found 6 transactions (324 active extents) in activity log.
> drbd0: drbdsetup [2573]: cstate Unconfigured --> StandAlone
> drbd0: drbdsetup [2586]: cstate StandAlone --> Unconnected
> drbd0: drbd0_receiver [2587]: cstate Unconnected --> WFConnection
> drbd0: drbd0_receiver [2587]: cstate WFConnection --> WFReportParams
> drbd0: Handshake successful: DRBD Network Protocol version 74
> drbd0: Connection established.
> drbd0: I am(S): 1:00000002:00000001:00000002:00000002:00
> drbd0: Peer(S): 1:00000006:00000004:00000027:00000006:00
ok, this time it is secondary/secondary,
so you just now forced it to sync your fresh made empty file system over
to the previously populated one.
your bad.
> Sep 10 14:33:03 ha1 kernel: drbd0: drbd0_receiver [2587]: cstate WFReportParams --> WFBitMapT
> Sep 10 14:33:03 ha1 kernel: drbd0: Secondary/Unknown --> Secondary/Secondary
> Sep 10 14:33:03 ha1 rc: Starting drbd: succeeded
> Sep 10 14:33:04 ha1 kernel: drbd0: drbd0_receiver [2587]: cstate WFBitMapT --> SyncTarget
> Sep 10 14:33:04 ha1 kernel: drbd0: Resync started as SyncTarget (need to sync 548737652 KB [137184413 bits set]).
> Sep 10 14:33:32 ha1 kernel: drbd0: Secondary/Secondary --> Primary/Secondary
too late, you made the already synctarget primary,
thus you are accessing the remote data.
--
: 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.