[DRBD-user] 8.4.0rc2 and volume tag

Gianluca Cecchi gianluca.cecchi at gmail.com
Mon Jun 20 13:02:01 CEST 2011


Hello,
is this "volume" tag that defines the "multiple replicated volumes in
a single DRBD connection" feature for 8.4.0rc2?

Suppose I have this in drbd.conf
resource "r0" {
  disk {
    on-io-error detach;
  }
  volume 0 {
        device  minor 0;
        disk            /dev/vdb1;
        meta-disk       internal;
  }

  volume 1 {
        device  minor 1;
        disk            /dev/vdc1;
        meta-disk       internal;
  }

  on drbdsrv1 {
    address     10.4.5.167:7789;
  }

  on drbdsrv2 {
    address    10.4.5.168:7789;
  }
}

is it correct to get something like this with drbd-overview command:
[root at drbdsrv1 ~]# drbd-overview
  0:r0             SyncSource Secondary/Secondary
UpToDate/Inconsistent C r-----
	[=========>..........] sync'ed: 51.7% (2030512/4194092)K
  1:??not-found??  SyncSource Secondary/Secondary
UpToDate/Inconsistent C r-----
	[=========>..........] sync'ed: 52.6% (1992496/4194092)K


This is on a CentOS 5.6 32bit + updates and drbd compiled from source
(it is a qemu/KVM vm on rh el 6.1 host) .

After the r0 resource is syncronized and drbdsrv1 is Primary, issuing
mkfs -t ext3 /dev/drbd0
generates a panic on it and drbdsrv2 going to:
[root at drbdsrv2 ~]# cat /proc/drbd
version: 8.4.0rc2 (api:1/proto:86-100)
GIT-hash: c0014a5ec7b162d3e4a3a81df829f9d4d84de94a build by
oper at jobsched, 2011-06-17 17:23:10
 0: cs:WFConnection ro:Secondary/Unknown ds:UpToDate/DUnknown C r-----
    ns:0 nr:4204576 dw:4204576 dr:0 al:0 bm:256 lo:0 pe:0 ua:0 ap:0
ep:1 wo:b oos:0
 1: cs:WFConnection ro:Secondary/Unknown ds:UpToDate/DUnknown C r-----
    ns:0 nr:4203820 dw:4203820 dr:0 al:0 bm:256 lo:0 pe:0 ua:0 ap:0
ep:1 wo:b oos:0

(at the moment only drbd installed form source and no clustering layer)

I got a stack trace... but not able to reproduce if enabling serial
console to cut and paste

(recorded video at
https://docs.google.com/leaf?id=0BwoPbcrMv8mvZjkzZjZmZDEtYThjOC00YThjLWJlYzgtNzBhZDk4YjhjNTA3&hl=it&authkey=CKuD0vIE)

BTW: during all the initial syncronization I got continuously these lines:
...
Jun 20 12:21:12 drbdsrv2 kernel: block drbd0: ASSERT(
drbd_interval_empty(&peer_req->i) ) in
/home/oper/rpmbuild/BUILD/drbd-8.4.0rc2/drbd/drbd_receiver.c:1739
Jun 20 12:21:12 drbdsrv2 kernel: block drbd0: ASSERT(
drbd_interval_empty(&peer_req->i) ) in
/home/oper/rpmbuild/BUILD/drbd-8.4.0rc2/drbd/drbd_receiver.c:410
Jun 20 12:21:12 drbdsrv2 kernel: block drbd0: Resync done (total 9
sec; paused 0 sec; 10668 K/sec)
not being able to access the console due to screen updates...

Gianluca



More information about the drbd-user mailing list