TOPIC:

Hello,

thanks for your reply.

We have a lvm on /dev/sda3: /dev/storage/lvstorage

So we configured disk /dev/storage/lvstorage in drbd.conf.

But our drbd is wating for connection...

drbdadm -f primary iscsi1:
Need access to UpTpDate data
drbdsetup 1 primary terminated with exit code 17

Thank you!

Guido

have your checked your replication network interfaces to ensure that each host can communicate with the other and the firewall is set correctly?

Yes, the network bond is working fine.
We´ve open the whole network as you´ve discribed in your howto.

Yesterday it was very late, so I forgot one thing that is also not really clear:

We build our servers with three NICs. Two NICs are bonded as you described and the third NIC is for the management interface and the bridge together.

We set FENCE_HEURISTICS_IPS to 192.168.1.254 which is the gateway ip for the brided interfaces.

Should this work correctly?!?

Thanks in advance

Guido

Hello again,

today I try it again with new installed systems:

Two XenServer 6.2 (XS62ESP1,XS62ESP1002,XS62E014) / 4 NICs, LSI RAID1
I prevent Xen from installing a "Virutal Machine Storage".
I create a "small" test partition /dev/sda3 (sda1 and sda2 are for xen dom0).
I use pvcreate /dev/sda3 then I reboot the system a last time and start with your howto.

in Step 7:
I ceate this LVM:
vgcreate vgroup /dev/sda3
lvcreate -L 40G -n lvolume vgroup

fdisk -l said /dev/dm-0 so I use "disk /dev/dm-0" in drbd.conf

drbdadm create-md iscsi1 gives me:
when I start the drbd daemon the counter runs but he didn´t find the peer node.

Then I stop the service and start again using "disk /dev/vgroup/lvolume;" parameter in drbd.conf.

After I initialized the Disks I start the service, but he didn´t find the peer node again.

I have checked my replication network interfaces.I can ping each host over his bond ip.

Any help would be greatly appreciated.

Thanks in advance

Guido

PS: Some system output:

gdisk output:


fdisk -l output:


/etc/tgt/targets.conf (host1)


/etc/tgt/targets.conf (host2)


/etc/drbd.conf (both hosts the same)

Hello Guido,
Regarding the fencing heuristics IP, yes, your chosen IP should be OK. We recommend traversing the management interface for this IP (which you have done)
By the way - this will have no impact on normal operation, it is only used when fencing a potential failed host. So,it is unrelated to the DRBD syncing issue.