DATACHECK errors with MSA1000 DGA Devices on OpenVMSV8.2

[Rob Leadbeater]

Hi,

If the firmware on the switches is too old, it might cause incompatibility issues with the newer firmware on the HBAs...although if the DS25's HBAs are also at the same firmware revision, that should rule that out...

I'm surprised that Storage Support haven't asked you to look at the switch port status, as that might indicate some faults, either with a HBA or cable etc.

Cheers,

Rob

[Joe Trimble]

The switches show firmware as follows. Both are identical, so only switch1 is displayed here.

MSA-VMS-1-switch1:admin> version
Kernel: 5.4
Fabric OS: v3.2.1b
Made on: Fri Jul 28 14:42:33 PDT 2006
Flash: Fri Jul 28 14:43:15 PDT 2006
BootProm: Mon Jul 8 18:35:44 PDT 2002

A quick check on the HP download site shows this is one minor revision behind on the fabric OS. The latest download is v3.2.1c, dated in 2007.

Joe

[Rob Leadbeater]

That version shouldn't cause any problems...

Does a "portErrShow" indicate any issues on the ES40's ports ?

Cheers,
Rob

[Joe Trimble]

Rob,

I'll attach a small file showing output from a few commands on both 2/8 switches, including porterrshow. I don't know how to interpret most of the stats.

I've sent the same information to HP Storage Support as part of my open call with them.

Note the ES40 is connected to port 1 on each switch; the DS25 is connected to port 2.

Thanks,
Joe

[Khairy]

hi joe,

can you provide the show tech_support output from msa controller. I know you've checked this but just want to understand more the configuration and the problem you have.

I've installed dual controller MSA1000 (14 x 72GB) with 2 x DS25 running openvms 7.3.2 last year without no problem. The only thing i encounter was patch issue. I upgraded both SRM to rule out firmware problem and it works for me.

> show tech_support

[Jon Pinkley]

Joe,

No guarantees, but have you tried a complete power cycle (complete removal of power cords) on the ES40?

We installed DS-A5132-AA (370426-B21) PCI-X 64BIT 133MHZ 2Gb-ALPHA LP10000, FCA2684 HBAs in ES40 M2 4 6/667 systems, and had problems until I did a firmware upgrade on the ES40, using the manual procedure and upgraded everything, including RMC, followed by a complete power cycle (per the firmware
upgrade instructions). For more details see this thread:

http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1140804

The storage was EVA on HP (Brocade) 4/16 switch fabrics. The error we saw was problems in autoconfig, not exactly the same as you are seeing, but I would still try, it only cost time.

Your porterrshow output looks fine. The "enc out" errors are related to auto-negotiation and they are normal. If you really dislike them, you can lock the speed, but in my opinion, that tends to come back to haunt in the future when you plug something else into that port (perhaps for debugging).

The MSA works with your DS25, which tends to point toward the ES40 as the problem location.

If you are going to need to power cycle, I would take the opportunity to do a firmware update on the ES40, and I would use the manual procedure to update everything, including RMC, followed by a complete power disconnect (not just turning off the power at the front panel. I shut my systems down, then disconnected the cords for a minute (perhaps overboard, but it isn't something you need to do frequently).

If that doesn't fix the problem, the next thing I would do is swap the FC HBA in the problem box. But I would try the firmware upgrade/power cycle first before introducing other changes.

Good luck,

Jon

[Joe Trimble]

I'm attaching a file containing the show tech_support command from controller 1, captured last Saturday.

Regarding ES40 firmware -- the latest firmware updates (7.3) were installed before the MSA1000 was connected. We needed to upgrade because our old firmware was not up to the correct level to recognize the MSA devices.

Current firmware revisions (taken from LFU capture last Saturday during a review):

Abios v5.71
SRM v7.3-1
pga0 TS1.91X6
pgb0 TS1.91X6
rmc V2.8
srom V2.22-G
tig 10

I have power-cycled the ES40 multiple times, but only by turning off the switch on the front panel. The LED's on the HBA's do go dark when the system is powered off. Do you think pulling the power cables will make a real difference?

Thanks,
Joe

[Jon Pinkley]

Is the firmware on the HBA's up to data?

The first time I did my firmware upgrade, I didn't cycle power and I didn't update the RMC (I hadn't used the manual update, which is still automated).

Whether it makes a difference? We have had several flakey problems on the ES40s (in this case a fan was being reported as bad, when it appeard to be working). A Field Service engineer replaced the fan, it worked, but the show power command continued to call out a bad fan. He was about to replace the motherboard (he had even ordered it), but before we did, he said he wanted to cycle power, with a complete disconnect.

That solved the problem.

So in at least those two cases, cycling power all completely, seems to have had an effect. The ES40 is never "completely off" when you use the fromt panel.

I think there is even something in the firmware update documentation that says you need to do a complete power cycle after appliying certain firmware updates (for example the RMC).

Like I said before, no guarantees. But I would try it. And if you haven't updated your RMC, I would do that too.

If you looked at the thread I referenced before, you will see we were getting errors. Have you analyzed your errlog with SEA?

Jon

[Jon Pinkley]

This thread is about a priblem that sounds exactly like the one we had that was solved by a complete power cycle. My guess is they really didn't need to replace anything.

http://forums.itrc.hp.com/service/forums/questionanswer.do?threadId=1146545

[Jon Pinkley]

Here's another recent thread involving ES40s.

http://forums.itrc.hp.com/service/forums/questionanswer.do?threadId=1247808

What "fixed" the problem was moving the CPU boards and memory banks. So the question remains, was it the complete power cycle (assuming they really did remove power when moving the modules), or was it the reseating of the components? If you don't try a complete power cycle before changing something, you never will know for sure what the fix was.

So I always do the complete power cycle first and verify that the problem still exists before starting to change anything else.

Jon