TOPIC:

Hi:

After a complete installed and running system i created 2 test VM with xencenter letting pool to decide where the VM will be booted and another one assigned to the slave machine:

- Slave DOWN: Everything work fine. Machines from slave are started on master.
- Master DOWN: After a While i connect XenCeter to the other host (now promoted to master) but the VM says VDI not available.

The SLAVE was promoted to MASTER: is this the HA-Lizard behaviour or is it a xen-pool-behaviour?

After the master is up again, the SLAVE remains as NEW POOL MASTER, but the iSCSI Storage is in XenCenter:
- "unplugged" state on old POOL MASTER and repair does not work.
- "connected" state on old POOL SLAVE (now the new POOL MASTER), but the machines belonging to the slave get theVDI not available message.

Is this the expected behaviour?

Note: iscsi-cfg status command shows the new master and slave roles changed as expected.

The slave being promoted to master is normal and required behvior of HA-Lizard as the SR will always follow the master.

Regarding the "VDI is not available error".

Can you check 2 things:

1 - ensure that there is no stale LVM cache on each host. Check if this file exitst (/etc/lvm/cache/.cache), and if so, delete it with:
rm –f /etc/lvm/cache/.cache && vgscan
when done, try running /etc/iscsi-ha/replug_pbd

2 - if that does not resolve the issue - take a look at this post for manually trying to resolve it:
halizard.com/forum/software-support/21-v...ble-on-second-server

I decided to remove and recreate again the iscsi1

service ha-lizard stop -w
service iscsi-ha stop -w
service drbd start
drbdadm detach iscsi1
drbdadm disconnect iscsi1
drbdadm wipe-up iscsi1
xe sr-list name-label=
xe pbd-unplug uuid=
xe sr-forget uuid=

dd if=/dev/zero bs=1M count=1 of=/dev/md0
drbdadm create-md iscsi1

service drbd start
drbdadm attach iscsi1
drbdadm syncer iscsi1
drbdadm connect iscsi1

service ha-lizard start -w
chkconfig iscsi-ha on
service iscsi-ha start -w

After testing again, when poweroff MASTER, SLAVE promotes to MASTER but i get the same VDI error, but now now, when the 2 mservers are up the status is:
DRBD CONNECTION: iscsi1 in StandAlone state

HOST ROLE:MASTER
DRBD ROLE:iscsi1=Primary
DRBD CONNECTION: iscsi1 in StandAlone state
ISCSI TARGET:tgtd (pid 18134 18132) is running...
VIRTUAL IP:10.10.10.3 is local

HOST ROLE:SLAVE
VIRTUAL IP:10.10.10.3 is not local
ISCSI TARGET:tgtd is stopped
DRBD ROLE:iscsi1=Secondary
DRBD CONNECTION: iscsi1 in StandAlone state


If i try:
drbdadm connect iscsi1

I get DRBD CONNECTION: iscsi1 in WFConnection state
And a moments later it returns to Standalone state

Any help?

Did you check your LVM filters and cache?

Also, watch the DRBD log while starting DRBD to see what is happening.. 'tail -f /var/log/messages | grep drbd'

Lastly - are you sure that iptables is set correctly? If unsure and this is a test system, try turning off the FW temporarily - 'service iptables stop'

The problem was a raid filesystem corruption.

Now everything works fine!

Thanks.

I ve got an SR in UNPLUGGED state:
I've tryed forget and reconnect SR (xencenter detects an does it but only partially. The first server is connected, but the slave is unplugged).




The VM's are running in both serevers (I cannot understand why if the SLAVE server has unplugged SR)!!!
*************************************************************************
HOST ROLE:MASTER
DRBD ROLE:iscsi1=Primary
DRBD CONNECTION:iscsi1 in Connected state
ISCSI TARGET:tgtd (pid 10980 10978) is running...
VIRTUAL IP:10.10.10.3 is local
*************************************************************************
HOST ROLE: SLAVE
VIRTUAL IP:10.10.10.3 is not local
ISCSI TARGET:tgtd is stopped
DRBD ROLE:iscsi1=Secondary
DRBD CONNECTION: iscsi1 in Connected state
*************************************************************************

I tryed a replug wathching th iSCSI uuid (i got 2 PBD iSCSI UUIDs with the same SR UUID, but i don't know if it is correct or no):
xe pbd-list sr-uuid=3d37a174-67c4-0126-416e-4ed168d58a00
uuid ( RO) : 6c95e7ec-e291-73d1-71a0-f03ac8e77a48
host-uuid ( RO): c6915619-6731-431f-ac50-357f17a7a1fd
sr-uuid ( RO): 3d37a174-67c4-0126-416e-4ed168d58a00
device-config (MRO): port: 3260; SCSIid: 1IET_0001000a; target: 10.10.10.1; targetIQN: iqn.2013-05.com.labs:x602-01
currently-attached ( RO): false

uuid ( RO) : fb55886f-c0b4-efc2-24c7-b75e78d7694c
host-uuid ( RO): c6a5d583-bc5a-4995-b474-3025e49f9e75
sr-uuid ( RO): 3d37a174-67c4-0126-416e-4ed168d58a00
device-config (MRO): port: 3260; SCSIid: 1IET_0001000a; target: 10.10.10.1; targetIQN: iqn.2013-05.com.labs:x602-01
currently-attached ( RO): true

Unplug works but i get an error while plugging:
xe pbd-plug uuid=6c95e7ec-e291-73d1-71a0-f03ac8e77a48
Error code: SR_BACKEND_FAILURE_47
Error parameters: , The SR is not available [opterr=ISCSI device failed to appear [opterr=during login]], (this doesn't work).
xe pbd-plug uuid=fb55886f-c0b4-efc2-24c7-b75e78d7694c (this one workrd)

vgscan shows NO Groups in SLAVE Machine (In master it shows a Group):
vgscan
Reading all physical volumes. This may take a while...
No volume groups found

Any help?