Symptom:
I have an ISCSI fail-over setup running on Open-E. The virtual interface
is on a bonding interface . On the storage are VM which are running in
Vmware ESX.
When starting three vm's on the hardware nodes storage becomes unstable
and we are getting the errors below in open-e.
Event group
2009/01/23 01:51:54
data_out_start(1032) unable to find scsi task 1d3f0 1788
2009/01/23 01:51:54
data_out_start(1032) unable to find scsi task 169d893 3465
2009/01/23 01:51:08
data_out_start(1032) unable to find scsi task 169d28c 2edf
2009/01/23 01:51:08
data_out_start(1032) unable to find scsi task 169d28e 2ee0
2009/01/23 01:50:46
data_out_start(1032) unable to find scsi task 1caa9 e55
Problem:
There is only one LUN and more than one initiator
connected. iSCSI Target works block based - not file based, meaning that changes
made by one person are not seen by another person working on the same
target/volume. Therefore unless you've got a cluster filesystem, concurrent writes
can cause trouble.
Solution:
Please changed the bonding from balance-rr to
active-backup and everything is working fine now
Additional information:
Open-E software impacted:
Example ticket number: