Re: Resp.: drbd WAS working fine; but now fsck can't open /dev/drbd? ??
Hi Lucas -
happy to share my little bit of knowledge :)
I've got 2 servers - a primary & a secondary. Each has 2 hard drives:
the first, /dev/sda, is a typical linux config'd drive. the second, /
dev/sdb, specifically /dev/sdb1 is setup to be my cloned partition
across the 2 systems.
I used package doc's, as well as some I found on howtoforge, to walk
thru setting up & configing the 2 systems. I was trying to figure out
why HA wasn't working correctly - I brought down the #1, and #2 saw &
acknkowledged this, but never mounted the sdb1 partition. The next
day, #1 would not boot up without complaining about fsck'g /dev/drbd0
(which is the drbd device for /dev/sdb1).
You can see in my last self-followup, where I'm at diagnosing the
problem.
not sure how much detail you'd like, but in case you do - here's my
config files: (lintest = #1; lintest2 = #2)
/etc/drbd.conf: ==>
resource varb {
protocol C;
incon-degr-cmd "echo '!DRBD! pri on incon-degr' | wall ; sleep 60 ;
halt -f";
startup {
wfc-timeout 10;
degr-wfc-timeout 30; # 2 minutes. -> 30 seconds
}
disk {
on-io-error detach;
}
net {
max-buffers 20000; # 2048 -> 20000 (play w/ this # for perf)
unplug-watermark 12000; # 128 -> 12000 (play w/ this # for
perf)
max-epoch-size 20000; # should equal max-buffers
}
syncer {
rate 10M;
group 1;
al-extents 257;
}
on lintest {
device /dev/drbd0;
disk /dev/sdb1;
address 192.168.1.59:7788;
meta-disk internal;
}
on lintest2 {
device /dev/drbd0;
disk /dev/sdb1;
address 192.168.1.58:7788;
meta-disk internal;
}
}
<==
on the HA side:
/etc/ha.d/authkeys:
auth 1
1 crc
/etc/ha.d/haresources:
lintest 192.168.1.55 Filesystem::/dev/drbd0::/mnt::ext3
/etc/ha.d/ha.cf: ==>
logfile /var/log/ha-log
logfacility local0
keepalive 2 # change to 1????
deadtime 30 # drop to 10????
warntime 10
initdead 120
udpport 694
bcast eth0 # Linux
auto_failback on
node lintest
node lintest2
respawn hacluster /usr/lib/heartbeat/ipfail
<==
On Jul 16, 3:30 pm, "Lucas Mocellin" <lucasmocel...@gmail.com> wrote:
> Hi Bob,
>
> I can't help you with your trouble, but I'm interested in drbd, can
> you post your cenario? I'll test it.
>
> Thanks,
>
> Lucas.
>
> 2008/7/16, Bob <bobg.h...@gmail.com>:
>
>
>
> > As a followup to my own post -
>
> > I've since tried inserting drbd into /etc/modules, and /etc/initramfs-
> > tools/modules (and updated initramfs).
>
> > But nothing I do seems to cause the drbd module to load prior to fsck
> > running..
> > (I do have drbd0.7-module-source installed) and the module DOES load
> > & work fine, if I manually continue past the boot issue.
>
> > Any other thoughts/suggestions?
>
> > TIA - Bob
>
> > On Jul 14, 1:00 pm, Bob <bobg.h...@gmail.com> wrote:
> >> running debian etch; drbd v0.7; along with HA; under ext3 filesystems;
>
> >> both system were running fine, and drbd work working great.
> >> assuming there was a power outage, now when I bring up the primary
> >> system I get (during boot):
>
> >> fsck.ext3: no such file or directory while trying to open /dev/drbd0
> >> the superblock could not be read or does not describe a correct ext2
> >> filesystem.
> >> fsck died with exit status 8 ....[snip]... superblock may be
> >> corrupt.
> >> <<
>
> >> /dev/drbd0 is drbd's device to /dev/sdb1; which DOES fsck PERFECTLY.
>
> >> this system isn't even coming up to runlevel 2... going into maint.
> >> mode....
>
> >> i've tried ef2check -f and it shows filesys clean... (again on /dev/
> >> sdb1)
>
> >> /dev/drbd0 doesn't exist until drbd is loaded; but if I load drbd,
> >> there's no way I can fsck /dev/drbd0 because "it's busy".
>
> >> anyone have any idea's ??
>
> >> TIA - Bob
>
Reply to: