"scsi downshift error"

[Andy_180]

A dual channel MSA30 connected to a DL 380 G3 via A 6404 a/c. Has anyone dealt with the error that pops up inexplicably in W2K sys log that says a fault has occurred in the expansion module of the controller in slot 1, drives are downshifting the speed. all F/W has been updated, different versions of SS have been used to built the setup. HP has replicated the issue in their labs ( I am direct contact with tier 3 support). It's been 2 months now with no resolve. when we see this error pop up, our DBs gets corrupt.

a new system including new Hd's all the way down to the RAM has been replaced. (including a new MSA 30)

HP says it doesn't cause corruption and not to worry. Microsoft PSS says they have 12 cases of other customers having the same issues and data loss. Hp says "no it doesn't". MS says "yes it does". i have been there and seen it, and I say "yes it does".

its a classic case of Godzilla's pointing fingers. in the mean time we are on pins and needles and are cutting more tape than a tape dispenser.

I'm looking for some insight of what others may have done to resolve this since HP cant get their own product fixed. Does anyone have this setup in production? what has been your experience?

thanks in advance!

--Andy

[JohnWRuffo]

I have seen this problem once before (outside of the bios fix one in 2.26) with an MSA... we traced it to a malfunctioning Power Supply in the MSA. Since the replace, I have not seen the message.

If you have replaced the whole MSA30, I wonder if this was a bad batch. Have tier 3 test on the power supply.

?

[JohnWRuffo]

btw... Yes... I am using your exact make up.

DL380 G3 with 6404 PCI Slot 1 (133)
MSA30 with a full load of drives 8x36GB and 6x146GB split across both channels.

[Andy_180]

thank John. blazing fast response. we have replaced the whole unit. we doa'ed it back to the vendor after a week of part swapping.

Also, we have 2 other idetical units that suffer the same fate. Hp is saying a F/W fix is being worked on but its been 2 months now. it seemed to come up the most after a warm boot but only 1 or 2 times per 10 reboots.

[Andy_180]

Let me give you some in depth config info. the 5i 2.58 is feeding the duplexed G3 backplane. 72GB 10K drives houses the OS, the other 4 in the G3 house the T-logs which have 72 GB 15k drives fed by channel 1A on the 6404. the MSA (DB house) is using 1B, 2B respectively and has 14 72 GB 15K Hd's. dual 3.2 Xeons, 2 GB DDR, all F/W is current and using the 7.1 agents. we've also tried every other agent down to 6.2, (that was my breaking point) . I've updated and back-reved the CPQISSM and CPQISSE ( base I/O driver (( miniport driver)) & the driver responsible for reporting physical health to the OS). w2k3 is not an option at this point in the game.

I dont see where the U320 Protocol is being comprimised despite the errors. It would be intersting to connect a scsi backplane analyzer and see for myself what is really happening.

[Gene Reese · Accepted solution]

Out of curiosity, has anyone progressed with this matter?

We are experiencing a similar issue, however I am inclined to believe that the downshifting issue is a symptom of some deeper, I/O problems with the hardware.

We randomly experience the same errors on roughly the same configuration as you did. The only viable theory I have stumbled upon was information regarding the lack of a battery backed write cache (BBSC) on the 5i controller. This, in turn, causes slow write speeds which may the source of the symptoms we are seeing.

Any additional information anyone could provide on this would be helpful.

Thanks!
-Gene

[Edmund White]

You need to make sure that the firmware on the MSA30 is up to date. It took some heavy work with HP reps to get the proper firmware (I think it was 2.30). There was no mention of the firmware on any HP sites, but the support person actually emailed me the Smart Start ROMPAQ. Contact me if you need more info.

[abcxyz]

Has this been resolved.

I have a similar issue and the firmware update did not resolved.

[Andy_180]

Yes, this was resolved but its been so long ago I don't recall exactly what eventually fixed it. it may have been a new version of the CPQISSE. also, remove the "hp diagnostics" from add/remove programs. for some reason i think that had something to do with the issue of scsi bus downshift errors.
thanks.
--Andy

[fengwang]

We have the same question. And the db corrupt happen after one month, we replace all hardware but motherboard.
hp dl760 6400scsi raid 5 four disk.
Only one step replace all four disk the error disappear.