EVA preventing from accesing discs

[GPissutto]

Hi everyone,
Last week I could not access disc services from my EVA8000. There were no warn's or fail's, it just happened that all the systems connected to this EVA didn't work, including the appliance. This appliance manages two EVAs (5k, 8k), and the SAN is connected through a 24k director. The appliance zone has these three items (app, eva5k, eva8k). I tried booting the appliance, but it didn't unless i disconencted the FC cable, same for other systems. By modifying the zones (removing eva8k) they were able to boot, but when I re-zoned with 8k again, they didn't work again. Well, after reseting one of the eva8k controllers everything got back to normal. But I still can't explain what happened.
The details are:
eva5k
Type: HSV110
Version: 3028
Software: CR0A3Evcsp-3028

eva8k
Type: HSV210
Version: 5031
Software: CR0B42xcsp-5031

Appliance
ProLiant DL380 G4, emulex fc card, CV 6.0

I´d appreciate any clue.
Could it be this appliance 3some zone?, fw?, fabric?, windows drivers?. If anyone has suffered this behavior, thanks very much since now.

[IBaltay]

Hi,
you should find some details on the issue in the EVA controller log

[IBaltay]

The other thing is, that both EVA 5000 and EVA8000 firmwares are obsolete/not supported and you should upgrade 3028 to 3110 and 5031 to 6220...

[GPissutto]

Hi IBaltay, thanks for your quick answer. The thing is that there is no error or warn reported neither in the controller log nor switch nor Windows. And I was suspecting FW, but I don't have anything reasonable to say to my boss that the problem is FW. Thanx again.

[Víctor Cespón]

There are many ways a EVA controller can hang, specially with an old firmware like 5031.

You'll need to capture the EVA controller event log and let us see...

[GPissutto]

The thing is that this controller was making others hunging up. Anyway, these are the logs:

06:50:59:992
2-May-2009
Controller A 04130003 #47286
Storage System event reporting is still active.
Corrective action code: 00 More details

18:32:19:815
1-May-2009
Controller A 04130003 #47285
Storage System event reporting is still active.
Corrective action code: 00 More details

06:22:07:031
1-May-2009
Controller B 0e070019 #47284
Description not found in parse file: HSV210_event_CR0B42_5031.txt
Corrective action code: 00 More details

02:20:12:416
1-May-2009
Controller B 0e070019 #47283
Description not found in parse file: HSV210_event_CR0B42_5031.txt
Corrective action code: 00 More details

01:13:42:018
1-May-2009
Controller A 0e030019 #47282
Description not found in parse file: HSV210_event_CR0B42_5031.txt
Corrective action code: 00 More details

22:16:07:480
30-Apr-2009
Controller B 0e070019 #47281
Description not found in parse file: HSV210_event_CR0B42_5031.txt
Corrective action code: 00 More details

The problem occurred last saturday (5/2).

znx very much.

[Víctor Cespón]

0e070019 -- The status of battery assembly '1' has changed.

0e030019 -- The status of battery assembly '0' has changed.

These are informational events, periodic tests of batteries. There must be something else on May 5, after 06:50:59.

[WilliamSmith11]

Some couple moth I had the same problem.
I find out that one GBIC have failed an also a disk drive have been failed too.

Send logs files

[Philip Immaraj]

Host port connection limit on B-series 3200 and 3800 switches
The EVA4x00/6x00/8x00 is limited to a maximum of three host ports on a single B-series 3200 and
3800 switch running version 3.2.x. HP recommends not exceeding more than one storage host port
connection on a single switch. If you are required to connect more than one storage host port to a single
affected switch, separate the connection into different quadrants. Connections are typically dropped
following an array controller resynchronization or when an event impacts the fabric, such as rebooting or
adding a switch.
Use the following options to avoid or manage the port limitation:
â ¢ For all hosts, zone by HBA as defined in the HP StorageWorks SAN design reference guide.
â ¢ Limit affected switches to only one HBA connection per host.
â ¢ Limit placement of the switch as an edge device and not part of the core.
If the switch drops a connection, reestablish as follows:
1. Disconnect the Fibre Channel cable from the failed port.
2. Wait 10 seconds and reconnect the cable. This will cause the port to relog into the fabric
and reestablish connection to the array.

To avoid this situation when running XCS 6.000 or earlier, set the C-series switch port to F-Port and
specify the speed (2Gb or 4Gb). Do not let the switch default to the Auto/Auto setting, which enables
the switch to negotiate the port setting and the operating speed.
If the switch does not log into the array, disconnect and then reconnect the FC cable on the array or
the switch.
If the above recommendations do not correct the problem, it will be necessary to restart the controller to
restore host port operation. You only need to restart the controller experiencing the host port problem.
This can be done from HP Command View EVA as follows:
1. Open HP Command View EVA and click the icon of the appropriate storage system. You can select
either an initialized or unintialized storage system.
2. On the Initialized Storage System Properties page, click Shut down.
3. Under Controller Shutdown, select the appropriate controller (A or B) for restart.
4. Click Restart.
The controller is restarted. After the restart, the host port should be operating normally.

[GPissutto]

There was a GBIC in one of the switches making all the noise.