Hello,
Summary: We experience a repeated boot problem, whereby we have to manually reconnect active Datastores on iSCSI targets.
Environment:
- 2 ESX 4.1 servers (on IBM X servers) all with latest firmware and updates.
- 1 QNAP TS-809U-RP NAS also with latest firmware, 8 * 2TB disks in RAID 6. We have 6 separate iSCSI targets with 1 LUN each of 1.9TB size.
- Connection between NAS and 2 servers is via dedicated Cisco/LinkSys gigabit SRW2016 switch, each box has 2 links active.
- 1 vCenter server controlling the 2 ESX servers.
Problem: Normally when all has been started, everything works fine, however when we reboot one of the ESX servers only the Datastores on the NAS (iSCSI) that are not being used by a Virtual Machine in the other active ESX server, will not come online (5 datastores at the moment).
I looked at logs and it seems the Datastores are seen as snapshots. Reconnecting these datastores via the vSphere Client does not work, because the only option presented is to format the datastore.
Work-around: The work around we have, which is manual unfortunately, is to connect to the console of the recently rebooted ESX server and use the command:
esxcfg-volume -M <VMFS UUID|label>
The interesting thing is that all 5 datastores reconnect in a few seconds by typing this command and the -M should make this persistent, but next time we reboot it happens again and it does not matter which ESX server reboots. Only when both ESX servers are rebooted at the same time and the datastores are not used by VMs the datastores reconnect automatically.
This problem has been with us for a long time, it also happened on ESX 3.5 and 4.0 and everytime with an upgrade/update we hope it is solved, alas not yet (maybe this is an inherited problem, however we remade the datastores 3 times now?).
Can you please help to solve this bugging issue?
Thanks, Edwin
