Smart array 3200 at Proliant ML530 said POST error 1783

[Ioanis Gaviotis]

4-year-old ML530 running W2K with Smart Array 3200 with 2 cages hosting 2 logical RAID-5 arrays each one on different SCSI channel. I simply rebooted it, having done no firmware upgrade whatsoever (System BIOS P19, SA-3200 firmware 4.50).

POST error 1783 appeared thrice. Then I reseated card, booted from SmartStart CD and it booted OK.

1. Is problem now solved? Should I reboot the server?
2. Why nothing was logged to IML?
3. During the problem, three out of 4 disk LEDs were out (not amber). Now all 4 are green. Could it be the cable connecting cage to controller?
4. Does the 3200 have any batteries that should be changed?
5. Does 1783 error appear only after reboot?
6. Available replacement for 3200 is 5302 (after 10 working days). Can I swap disk arrays from 3200 to 5302? And then back from 5302 to 3200 (if need be)?

Thanks,
Ioanis

[JohnWRuffo]

Ioanis:

A 1783 Controller failure error that shows up out of the blue is usually not good.

You can try to "fix" the error with re-flashing the bios.

A battery discharged error would be a 1793 or 1794. Unless this accompanied the controller failure error, it is not lilely the battery at fault (though I am sure they are well past their life expectancy; 3 years usually)

Yes, you can migrate to the 5300 series controller. Here is a document that verifies the migration path from the 3200 Controller:
http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?locale=en_US&taskId=115&prodSeriesId=374791&prodTypeId=329290&objectID=c00180566

[Ioanis Gaviotis]

Thanks John,
1783 is not good for sure.

Since no pre-failure warnings were given, no problem was logged, and there is no check procedure, how do I know if I actually need to order a replacement controller?

I mean, I can't even reboot the server to see if the 1783 will happen again, because that would cause the server to go down for good!

Is it valid to say that for any production server one should order two storage controllers in the first place?

[JohnWRuffo]

Having a spares around for critical 24/7 servers (or application servers needing three 9's or better) is a good idea.

I must say, though, that you are going to have a harder time getting "fresh" SA3200 controllers, except through HP. They will be installing new battery packs on the cards. Any third party vendor I would not trust to have "unused" product. Remember, the lifespan begins at first charge; then three years usually after that.

Yes, If I were you, I would replace it ASAP and get a good backup soon.

[Anders Norrbring]

I've gotten this now as well, when the controller has initialized itself it posts the following to the screen:

1783-Slot 2 Array Controller Failure!
[Command failure (cmd=11h, err=020h)]

I have no idea whatsoever what these codes are trying to tell me.

Anyone, please?
Anders.

[Ioanis Gaviotis]

Anders,
in short this error means that your SmartArray controller does not function anymore. I am really curious what could go wrong with a piece of equipment that has no moving parts and has no problem with electrical supply!

I replaced mine with a new one I ordered from the spares of HP. It is spectacular that HP Hellas had a spare SA-3200 readily available and sent it the same day I ordered it.

Advice: if 1783 appears during the boot of your server, don't panic. Try re-booting it after some time. In my case, (thank God) the server finally booted once without 1783 and ran OK, until the replacement came. During experimentation, I was able to repeat this phenomenon (intermittent 1783's with normal operation when boot without 1783).

Ioanis

[Anders Norrbring]

Alright... I've seen quite some threads about the 3200 and this error, it seems like a "family-sickness".. ;)

I've tried to re-seat the cache daughter card, I've also pulled out all the socketed circuit chips and re-seated them, still the same error. I've also tried to move the controller to several different machines, both Compaq's and others. No luck.

I guess it's time to give up and shop for new controllers, I have 3 of these "1783" boards...

Later,
Anders.

[Ioanis Gaviotis]

Ups, 1.5 years after substituting the failing SmartArray 3200, with a new one ordered directly from HP, the latter failed, too. With the same POST error 1783.

Will HP find me a replacement this time?