Software-based VM-centric and flash-friendly VM storage + free version
Moderators: anton (staff), art (staff), Max (staff), Anatoly (staff)
-
CedricT
- Posts: 36
- Joined: Mon Apr 15, 2019 1:14 pm
Thu Apr 25, 2019 7:34 am
Hi all!
We have an issue with our lab of vsan free with 2 nodes Windows Server 2016.
These nodes are directly connected throught 10GB connections. I mean for SYNC and ISCSI they are directely connected (no switch between them)
So here what happened:
We stopped one node (the second node) => the first node is still working great
We stopped the other node (the first node)
We started the first node => impossible to get the volume back with iscsi because network interfaces 10GB were considered down
We started the second node => the iscsci volumes came back immediately
So, is there any way to fix this problem?
-
Oleg(staff)
- Staff
- Posts: 568
- Joined: Fri Nov 24, 2017 7:52 am
Thu Apr 25, 2019 12:27 pm
Hi CedricT,
Could you please give more details? Are you using Loopback interface for ISCSI connections on both nodes? If no, please connect
-
CedricT
- Posts: 36
- Joined: Mon Apr 15, 2019 1:14 pm
Thu Apr 25, 2019 3:18 pm
Hi,
Thank you for the reply
Yes we have configured loopback already.
With iscsicpl we can see targets are inactives and we can't connect them.
If i try with targeting loopback address manually that still doesn't work.
What details do you need? I follow your guide for the configuration of 2 nodes with Hyper-v server 2016.
Thx you again
-
Oleg(staff)
- Staff
- Posts: 568
- Joined: Fri Nov 24, 2017 7:52 am
Thu Apr 25, 2019 4:11 pm
Hi CedricT,
With iscsicpl we can see targets are inactives and we can't connect them.
Did you try refresh option before connecting?
-
Boris (staff)
- Staff
- Posts: 805
- Joined: Fri Jul 28, 2017 8:18 am
Sun Apr 28, 2019 4:12 am
We stopped one node (the second node) => the first node is still working great
We stopped the other node (the first node)
We started the first node => impossible to get the volume back with iscsi because network interfaces 10GB were considered down
We started the second node => the iscsci volumes came back immediately
The above description is good, but unfortunately I seem to be missing part of information. Does the issue persist after the second node is up?
If the issue was happening only when node 1 was already up, that is something you would expect. It would need to communicate with the partner server to make its disks accessible for iSCSI clients, even though it was powered off the latest and powered on the earliest - it does not have that confirmation until its partner is accessible. This check is required to ensure data integrity.
-
CedricT
- Posts: 36
- Joined: Mon Apr 15, 2019 1:14 pm
Mon Apr 29, 2019 7:52 am
Hello,
Thank you for your detailed reply.
I want to be sure i understood this point.
For example, let's imagine we have a complete power off. Both servers are shutdown badly. One is completely down and can't restart. I restart the other one but it can't provide any access if i follow your explanation.
Is that right?
Not saying it's bad, i just want to know the limits and which case can be a problem.
-
Boris (staff)
- Staff
- Posts: 805
- Joined: Fri Jul 28, 2017 8:18 am
Mon Apr 29, 2019 3:32 pm
In that case, you would need to manually mark the disks on the live node as synchronized. This will make the disks accessible for the iSCSI initiators and you will be able to access your data.