Problem with MSA1000 and bl35p

[selpa_1]

Hi all

I have Blade BL35p servers connected to a MSA1000 SAN

I have upgraded all firmwares to the latest version.

Everytime I do a shutdown of a blade server, it issues a bus reset to the SAN and the active controler becomes passive. Sometimes the controlers start bouncing between Active and passive. And the MSA 1000 controlers fail.

I have been working in this issue over the phone with HP tecnichian over an week, with no sucesses.

As anyone experienced similar problem? Do you have any sugestions?

Thank you in advance
Selpa

[The Spartan]

HBA make,firmware and driver?
HBA to MSA connectivity - ie is there a switch/hub ?
OS on the blade?
PSP on the blade?
multipathing s/w in use?

[Uwe Zessin]

Does this advisory help?
http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?objectID=c00744940

[selpa_1]

Hi

Thank you for your answers

The Spartan:

Here is my current configuration

HBA QLOGIC Mezanine in all Bl35p - FW: 3.03.01 , BIOS Version 1.45, Driver: v9.1.0.18 (Boot from SAN Driver - Storport - latest)

MSA 1000 with 2 controllers in Active/Passive configuration - FW: v4.48 (latest)

2 x External Switches 2/8 - FW: v3.2.1b (latest). Zoning is active on both switches.

BL Enclosure G2 - FW: 2.30

BL Power Enclousure G1 - FW: 2.30

Each Blade has a dual connection witch connects to each switch. The MSA is also connected to both switches.

Operating system is Windows 2003 Enterprise R2 and Windows 2003 Server R2. The SW used for failover is MPIO Basic v1.3. PSP Version 7.60


Uwe Zessin. Thanks for the tip, but those are not our simptoms. The Server Blades, boot correctly from the SAN. The problem is that when several OS images are being deployed at the same time, or when some Server Blades are shut down, the MSA Active controler becomes passive, (and vice-versa) and sometimes they start bouncing between each other.

Because the blade can only boot from one specific MSA controler, after the controller swap, servers can't boot anymore.

Thank you for any help
Selpa

[The Spartan]

Hi Selpa,

Shall have a discussion with my colleagues and let you know of any info...

Sorry, am clueless presently..

[Jake Barber]

Just yesterday the same thing happened to me.

Heres the configuration:

MSA 1500 w/ MSA 30
2 - Storage Works SAN switch 2/16v (4.4.0b)

We have 5 servers that are plugged into the switch fabric.
DL360 G2 (Windows 2000)
DL360 G4 (Windows 2003) (Secure Path)
DL385 G4 (Windows 2003 x64bit) (Secure Path)
DL580 (Windows 2003) (Secure Path)
ML570 G1 (Windows 2000 - Old server, being retired) (Secure Path).

Scenario:

We were migrating the ML570 over to the DL385 (SQL). So we remapped the LUN that was going to the ML570 over to the DL385. Everything was fine, systems up and running for over a week. No issues.

Yesterday, I issued a shutdown of the ML570, since I was going to remove it from the rack. When the system shutdown, the entire Array failed over to the passive controller. The other servers in the array failed over and were able to continue using their assigned LUN.

We've been using the SAN for over a year, during that time I've had to reboot the DL360 G4 and the ML580 and things rolled right along without issue. But for some reason, shutting down the ML570 caused the Array to failover.

The ML570 has the following HBAs:

2 - FCA 2101's (5-5.10a11 SCSIport driver)

Any help someone could provide would be appreciated. I'd love to get this server out of here. Should I disconnect the fibre cards from the switch before issuing the shutdown of the server?