Event id 9 (kgpsa timeout) causes W2K to crash

[Eugeny Brychkov]

Ricardo,
here's the number to call
808 208 113 option 3 (local call)
try it, or email 'Suporte.Portugal@hp.com'
Eugeny

[BR765533]

Hi Eugeny,

I've talked with the support to elevate the situation to the competency center, which I was told had already been done. The situation is being evaluated by the Ireland Competency Center.
In case you have time, I'm attaching all the info I have at the moment, and I post whatever else I remember, will yet gather or info you might deem necessary, so you can take a look at it.

Thanks for all your help and availability.

Ricardo Sousa

[BR765533]

Eugeny,

You will find in the attached file some more info, that i've consolidated.

If you think you may need more data, please ask.

Ricardo Sousa

[Eugeny Brychkov]

My colleague who is more expert in your equipment than I am is dealing with it.
I've took a quick look and feel that you've upgraded drivers, but did not upgrade FCA2101 firmware and bootware - why? In addition what I see is that ID9 events are generated only for segment to which MSL is connected...
Eugeny

[BR765533]

Hi,

I conclude from your remarks that you have spoken with your colleagues in charge of this situation. Have you shared the case? Have you reached any new conclusions or theories aout the problem?
Regarding your other remarks: no I haven't upgraded the drivers. From the start the HBA were installed with kgpsa driver version 5-4.82a14. I've upgraded only a couple of servers with kgpsa driver version 5-4.82a16, and that was on a early stage of my attempts to solve the situation, as have mention it in the postings. I didn't upgrade the firmware because I read the release notes, and there was nothing refering to a possible solution to my problem, and because on other threads I've read, dated much earlier than my that particular version of firmware wasn't yet released, although i didn't rule out that possibility, I simply haven't come to that yet. Do you think that may be the solution?

And yes you are right, the timeouts only occur on the MSL Segment. I thought that had mention it before. That a fact I noticed from the start of the problem. More, the situation only started after the arrival and subsequent connection of the MSL to the SAN.

To happen to have any new thought or idea to share?

Again I would like to thank you for your help, time and availability.

Ricardo Sousa

[CA1034283]

I have msa 1000 with 3-dl360-g3 and 1-sp750 workstation with fca2101's. First warning is don't upgrade the firmware.... I upgraded and had to replace all hba's to get backup to run again, using Backup exec v9 and tried v8.6. I tried all the setting changes and havn't seen much change. I also tried zoning vs no zoning and seen very little difference. I've call support 7 or 8 times over 2 months with some help. The latest suggestion was to set the cache to 0% read and 100% write(10/9/2003). Support blamed sp3 on sql as the reason for that. My 3 servers average a blue screen every 3 days. Havn't had server crash since changed the setting 4 days ago new record.

[BR765533]

My problem seem to have disappeared. The only changes made so far included the following:
1. Disable Fibre Information Agents in Insight Manager
2. The port 0 on both FC Switches showed huge amounts of errors. The SAN Management Appliance was/is connected on that port. Since the FC HBA for SMA are 1Gb, the default value for that port (negotiate) was changed to a fixed 1Gb.

It´s not absolutetly clear which one of the changes did the trick, but the fact is the errors are gone, and my customer is now running a stable SAN enviroment.
There is still some questions in the air, and I know for a fact that HP support is looking into them, but those are minor "problems" and fine tune issues.

I'll post any evolutions as they arise.

I would like to thank HP support and Eugeny in particular for all help in troubleshooting and solving this issue.
Ricardo Sou

[CA1044853]

Hello all together,

we did have the same situation. We have a MSA1000 with 8Port Borcade Switches, Securpath 3.0B for Netware and 4.0C for Windwos. We use FCAL2101 for 4 Windows Servers, where we used FCAL2112 NetWare

We expirienced the exact same situation, and our solution was, to change all SwitchPorts from Negotation "N" to fix "2G" whitch solved the whole Problem.

About the Other Settings, found in this case, we have:

Driver Version: 5-4.82a16
BIOS Emulex:
RESET_TPRLO: 2
EmulexOption: 0xBA00

(All the above was set by default, no changes from our side)

Thanks for all the usefull Information from your side

Oliver