StarWind Software

Hello,

i have a new HA cluster with 2 hyper-v hosts with 3 disks

1 whitness 100MB
1 VOL1 1,5TB
1 VOL2 1,5TB

nic's are direct attached as folows

SYNC HOST1 10gbit nic direct attached to 10gbit nic host 2 (no redundancy, yet)
host 1 sync ip 10.0.100.10
host 2 sync ip 10.0.100.20

HB HOST1 10gbit nic direct attached to 10gbit nic host 2 (no redundancy, yet)
host 1 sync ip 10.0.101.10
host 2 sync ip 10.0.101.20

all iscsi targets are HA iscsi attached with 2 MPIO connection (exept for the whitness which is only locally connected). All of the VM's were registered on the host i would NOT reboot.

the problem is when i reboot 1 of the 2 hosts the storage goes completly offline, so there is no failover. i don't know of this is normal. It was also a "clean" reboot so the hosts could have negotiated.

attatches you find a few screenshots of my setup and the events of the host which remained online (so not the host i rebooted) and the storage went offline.

the storage came back up pritty quickly after the reboot, but this is not HA as it should be offcourse.

Hi,

Welcome to StarWind Forum. First, what is the network mask for the sync? If it is /24, then Sync adapters are in different subnets.
Also, it is expected for the host to go out of sync on restart. Say, if you restart host 1, host 2 will report host 1 out of sync, Once the server is up, host 1 should have fast sync. There can be a full sync too if HAs have write-back cache and the server was shut down unexpectedly. Let me know if you have more questions.

Did storage of the partner go offline (i.e., not synchronized) too? Was you have storage availability compromised on the active node?

yaroslav (staff) wrote:Hi,

Welcome to StarWind Forum. First, what is the network mask for the sync? If it is /24, then Sync adapters are in different subnets.
Also, it is expected for the host to go out of sync on restart. Say, if you restart host 1, host 2 will report host 1 out of sync, Once the server is up, host 1 should have fast sync. There can be a full sync too if HAs have write-back cache and the server was shut down unexpectedly. Let me know if you have more questions.

Did storage of the partner go offline (i.e., not synchronized) too? Was you have storage availability compromised on the active node?

how do you mean the sync adapters are in different subnets?

host 1 sync ip 10.0.100.10/24
host 2 sync ip 10.0.100.20/24

seems to be the same subnet not ?

i know that the sync is expected to go out of sync while restarting 1 node, but the storage on the remaining node goes offline which is not expected behavioure , the question is, why does the storage goes offline?

Got it. Please collect the logs from both hosts as described here https://knowledgebase.starwindsoftware. ... collector/ and share them with me.

yaroslav (staff) wrote:Got it. Please collect the logs from both hosts as described here https://knowledgebase.starwindsoftware. ... collector/ and share them with me.

just sent you a PM

Thanks!
I need logs from the 2nd server too. Collect them and share with me, please.

done

Did the log check. The problem was that you did not let sync to finish prior to restart. 01 was still doing sync when 02 was restarted. See the restart procedure here https://knowledgebase.starwindsoftware. ... installed/.

If the problem repeats even with devices synchronized on both sides, send an email to us support@starwind.com