POST Error: 1786-Drive Array Recovery Needed

[AndreasM]

Hi,
On a Proliant DL580 G2 I get the message
POST Error: 1786-Drive Array Recovery Needed
in Compaq Insight Manager 5.5.
The light is yellow for degraded.
On the server in the HP System Management Homepage it says for two drives:
Status: Predictive Failure
Action: Replace S.M.A.R.T. Drive

First Drive:
Under the heading Failure Indicators:
Everything 0 just
Hardware Errors: 6

Under Heading Statistics:
Hard Read Errors: 432
Recov Read Errors: 153

A second drive on this same server shows this:
Heading Failure Indicators:
Not Ready Errors: 6
Everything else under this Header is 0 (so OK).

Statistics (of second degraded drive):
Hard Read Errors: 13
Seek Errors: 4
Hard Write Errors: 1
Recov Write Errors: 7
Everything else is 0 here.

The server works just fine.
I believe we had the yellow degraded status already some other time and just marked it as repaired in Compaq Insight manager.

My feeling tells me that the disks actually are still ok and do not need to be replaced yet.
Can I get the system to just get these faulty spots repaired or marked that it doesn't use them? Like with chkdsk /F on small computer drives under windows?
What happens if I just mark it repaired in Compaq Insight Manager. Would that be dangerous?
Or what could I do?

Thanks for answering.
Sincerely
Andreas

[Johan Guldmyr]

Heya

some disks are prematurely being marked as predictive failed. This have been fixed in some firmwares.

Don't know about marking them as repaired in CIM. About how to make the disks not use those sectors, supposedly this depends on what error they have seen (maybe it's a mechanical error?). I thought disks stopped using those sectors by itself.

[marcus1234]

some drives do show false posotive" but Hard Write Errors are the "real" give away that particular drives needs to be replaced, especially if an amber light is lit on the disk ,but the diode can fail so pay attention to the hardwrite error in the logs..

chkdsk etc only mark the drive for the OS index , the array controller does not make use of this marker ,

I advise replace this drive because if you do not you could end up with parity errors on both drives and a broken array,
and i reckon this is a mirror and to lose both is not fun unless you have a sold backup

Enjoy
points appreciated :)

[TarunJain]

Hi Andreas,

There are many advisories recommending to update firmware of hard drives and Smart Array Controllers, to stop excess errors and to correctly identify when hard drive actually needs to be replaced.
Similar to this one:
http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?lang=en&cc=us&taskId=125&prodSeriesId=2510357&prodTypeId=3709945&objectID=c00792595

However in you case, The Drive Array is already in degraded state therefore The data is already at risk.
If you really want to use these hard drives a little more then better you should confirm their health by following steps:
1. Make sure to backup your data
2. Update firmware of controller and hard drives
3. Check the status of hard drives using ADU (Array Diagnostics utility)
4. If the error for hard drives is still present then replace your hard drives ASAP

[marcus1234]

I would certainly not update any firmware on a drive that showed hard write errors " , as drive internal controller could already be corrupt, and new firmware may exasperate the issue.. hP wouldcertainly change these disks as soon as you mentioned hard write errors and the amber led, and thats for a good reason..:)