No BUS reset possible - Starw. not 100% MSclustercompatible?

[ThomasS]

Hi forum,

Test Environment:
2 Node Cluster with 2 DELL PE Server (3 NICs: private, public and iSCSI)
W2k3 SP2 + all updates; Microsoft iSCSI Software Initiator 2.07
Starwind 3.5.5 Target: 1 NTFS Quorum (flat file) with cluster option with 3 NICs: (10.192.120.100 to Node1 iSCSI and 10.192.122.100 to Node2 iSCSI)
I've read your step-by-step configuration guide for ms cluster.

All my cluster tests are OK (Hardreset, Failover ..), but one TEST faild:

Node1 and Node2 are alive, Node1 is online (have all ressources)
and I cut off the private und public nic link connection for test

Result:
Node2 try go online and try to get the quorum ressource (but node1 have the quorum, but have no communication to node2) and Node2 send a bus reset scsi signal, but this failed with:

Eventlog Error:

Event ID: 1209
Description: Cluster service is requesting a bus reset for device \Device\ClusDisk0.

starwindlog:

..
LUN 0 is SR reserved by iqn.1991-05.com.microsoft:itcl02.domaene2.de,400001370001
..
S[4] ScsiOp 0x00 - Reservation conflict.
..
Std: S[4] ScsiOp 0x28 - Reservation conflict.
..


Node2 tries 3 times to reset the bus and then stoped the cluster service.
The cluster is down.

Is this a bug or a feature?

I think: Node2 must send the bus reset, then all nodes release the iSCSI Target and then Node2 get the quorum.

I am right?

[ThomasS]

An hard shutdown works in my config,

but nevertheless I tried

Try to change the following parameter in StarWind.cfg and restart the service:
from <iScsiPingPeriod value="0"/>
to <iScsiPingPeriod value="5"/>


http://www.starwindsoftware.com/forums/ ... c1d14dab47


But it don't works.



Greatings Thomas

[Bohdan (staff)]

You are building the configuration like in the following article
http://technet.microsoft.com/en-us/libr ... 27114.aspx
Right?
Have you configured networks for cluster use?
http://technet.microsoft.com/en-us/libr ... 28293.aspx

Anyway by cutting off both the private and public NIC link connections the cluster nodes have no ability to communicate and therefore node2 is trying became active and to take the clusters resources while they are in use by the node1.

If your StarWind version is 3.5.5 set the
<iScsiPingPeriod value="0"/>
in 3.5.4 version you should set
<iScsiPingPeriod value="5"/>
But it is better to update to 3.5.5 because it has many fixes.

[Val (staff)]

Hi,

The iSCSI quorum disk is the most critical resource of the cluster.
So I'd recommend to use only 2 NICs on the hosts: public and private+iSCSI.
This should remove the problem you described.

[ThomasS]

You are building the configuration like in the following article
http://technet.microsoft.com/en-us/libr ... 27114.aspx
Right?

I take this (W2k3 Server)
http://technet.microsoft.com/en-us/libr ... 83714.aspx

Have you configured networks for cluster use?
http://technet.microsoft.com/en-us/libr ... 28293.aspx

YES

Anyway by cutting off both the private and public NIC link connections the cluster nodes have no ability to communicate and therefore node2 is trying became active and to take the clusters resources while they are in use by the node1.

But node2 send an bus reset!


If your StarWind version is 3.5.5 set the
<iScsiPingPeriod value="0"/>
in 3.5.4 version you should set
<iScsiPingPeriod value="5"/>
But it is better to update to 3.5.5 because it has many fixes.

I've StarWind 3.5.5



I don't agree. In W2k3 cluster a online node1 lost all communication, than node2 with communication to public and AD-Controller try to take the onership of Quorum.

An that happens: In Eventlog node2 initiate an SCSI BUS Reset, but StarWind shows no reaction (I displayed the starwind log with tail -f).
After a bus reset, node1 must release the Quorum and node2 take it.

[ThomasS]

I've found:


How the Cluster service reserves a disk and brings a disk online
http://support.microsoft.com/?scid=kb%3 ... &x=11&y=11




Each node in the cluster renews reservations for any LUNs it owns every three seconds. If the nodes of a cluster lose network communication with each other (for example, if there is no communication over the private or public network), the nodes begin a process known as arbitration to determine ownership of the quorum disk. The node that wins ownership of the quorum disk resources in total communication loss between cluster node will remain functional. Any nodes that cannot communicate and cannot maintain or acquire ownership of the quorum disk will terminate the cluster service and any resources that node was hosting will be moved to another node in the cluster.
1. The node that currently owns the quorum disk is the defending node. The defender assumes that it is defending against any cluster nodes that it cannot communicate with and for which it did not receive a shutdown notification. The defender continually renews its reservation to the quorum by requesting a SCSI reserve be placed on the LUN every three seconds.
2. All other nodes (nodes that do not own the quorum disk and cannot communicate with the node that owns the quorum resource) become challenging nodes.
3. When the challenger detects the loss of all communications, it immediately requests a bus-wide SCSI reset to break any existing reservations.
4. Seven seconds after the SCSI reset requested, the challenger tries to reserve the quorum disk. If the defender node is online and functioning, it will have already reserved the quorum disk as it typically does every three seconds. The challenger detects that it cannot reserve the quorum, and terminates the cluster service. If the defender is not functioning properly, the challenger can successfully reserve the quorum disk. After ten seconds, the challenger brings the quorum online and takes ownership of all resources in the cluster. If the defending node loses ownership of the quorum device, then the cluster service on the defending node terminates immediately.
When a cluster node takes a disk resource offline, it requests that the SCSI reserve be released and then the drive will once again be unavailable to the operating system. Anytime a disk resource is offline in a cluster, the volume that the resource points to (the disk with the matching signature) will be inaccessible to the operating system on any of the cluster nodes.

[Val (staff)]

ThomasS,

Could you zip the StarWind Service's logs and send them to my email?

[ThomasS]

yes

[anton (staff)]

Nothing arrived yet Did you send it already?

yes

[ThomasS]

Nothing arrived yet Did you send it already?

yes

I will do it now.

[ThomasS]

The logs are on the way.

[anton (staff)]

Got 'em, thank you!