Netfinity 5100 - All 6 SCSI drives defunct 2

[LesterDK]

Hello techies

I'm currently stuck with a Netfinity 5100 server (8658-41Y) that gives the error:

1962 Drive does not contain a valid boot record.

This makes sence since all 6 disks are defunct according to raidstatus. (screenshot)

Raid setup
(ServeRaid-4M)
(All IBM 18.2GB 10k rpm disks)
1: c: system drive
2: Mirror of disk 1
3: d: data drive
4: Merged with 3 as 1 volume
5: Mirror of 3/4
6: Mirror of 3/4

Additional info
1. The server is/should be running Windows 2003 server.
2. Tried running diagnostics during startup. Pretty sure the controller passed the test. It could not find any drives though.
3. Using the controller config it only detects ID: 7 as being the controller itself.
4. I can follow instructions given - I don't got much know-how in this area.
5. Have not tried anything on the raid yet, since experience tells me this is a bad idea before knowing what to do.
6. No additional disks awailable at this time.
7. NO backup of the data of these disks. DOH

Hope someone will be able to help me out here.

/Kenneth

 

[ibmtech65]

Get yourself a server raid SUPPORT CD and boot to it. Do not do any updates if asked.
Right click on local system and in SERVRAID ACTION, save printable configuration and EVENT LOGS. (will save to a manually formated floppy disk).
Open txt file on another machine and at the bottom will be the order the drives went down.
You can force the drives back online in reverse order by right clicking on the drive in raidmanager and set state online.
You will not be able to set the last drive in an array online. Other words if you have Array A with 2 disks you can only set one online or if Array A was 5 disks you could do 4 of the 5.....
Once you have the drives online you should be able to boot to the OS.
If everything in the OS looks good, ie you can see all your data. You can rebuild the arrays simply pull one of the defunct drives, should have amber LED on front, wait a few minutes and put the drive back in. It should automatically start rebuilding. You can only do one rebuild at a time.

Also if you have a 4m card you might want to read this and consider disconnecting the battery on your raid card if the machine is not covered by some sort of maintanence agreement where you could get the battery replaced.

http://www-1.ibm.com/support/docview.wss?uid=psg1MIGR-61088

 

[LesterDK]

Thanks for your reply.

So far it seems pretty straightforward, and I will try the solution when I get hold of a floppy-disk.

I phoned IBM DK and got a technician who said more or less the same. Regarding the battery, he said I should just buy a UPS.. lol.

Will return with the results.

/Kenneth

 

[LesterDK]

Got it going again now.. 5/6 drives back online and looks fine so far.

Quote: You will not be able to set the last drive in an array online. Other words if you have Array A with 2 disks you can only set one online or if Array A was 5 disks you could do 4 of the 5.....

So what do I do with the last drive thats still defunct? I got both logical drives up. ATM disk 3 is the one still Defunct.

Will I have to get a new disk for that last spot, or will that do me no good?

I choose to make a backup of all data to another PC in the network using the Server Management Console.

 

[ibmtech65]

Since thet still defunct drives are part of the raid array simple remove the drive with the machine up and running on the OS, wait a few minutes and then put it back in, it should start rebuilding.
As far as the battery goes a UPS is totally different. The battery on the card is for the backup cache. The controller writes the data to the battery first before the hard drives that way should there be a unplanned power lose you don't lose data.
Since these batterys are near end of life they will start to cause problems if not replaced or removed from the raid card.

 

[LesterDK]

Did what you said... both LEDS light up now.. but cannot see any indication of it doing rebuild..

Installed ServRaid windows software aswell.... tried choosing rebuild from there.. but get an error doing it.

How should I be able to confirm that rebuilding should be in progress?

 

[ibmtech65]

Solid green light, flashing amber means rebuilding.

 

[LesterDK]

hmm.. both LEDs just stay on... amber one does not blink... been like this for hours now...

So for me pulling it out.. and putting it in does nothing it seems.

 

[LesterDK]

One of the time I tried this... the other disk in the spanned drive also failed and got defunct again.

Got it online with no problems, tried once again from inside ServeRaid Manager to set disk 3 online. Still getting an error that tells me it cannot be put online.

I'm guessing that this drive is gone with the wind. Thinking I will buy 1 or maybe 2 new drives just have a spare aswell.

Cannot find a compability list anywhere though, but found part.no: 90P1305

IBM harddisk - 73.4 GB - Ultra320 SCSI
Type: Harddisk - hot-swap
Model: 3.5"
Capacity: 73.4 GB
Interface: Ultra320 SCSI RaidServe 4M is 160 - problem?
Connector: 80-pin Centronics (SCA-2) Not really sure here. But assume I can re-use the casing from the drive I'm replacing (IBM Ultrastar Model: DDYS-T18350)
Speed: 320 MBps
RPM: 10000 rpm
Buffer size: 8 MB

Could not find any of the "old" disks and my experience tells me they are actually more expensive than newer/bigger models.

The only issue now is regarding it going to work or not. Maybe you got a better recommendation?

 

[LesterDK]

Just realised from looking more closely to the old IBM disk that it's 80P LVD/SE.

Really hoping I can buy a ultra320 and that it then supports 160... seems like quite a pain to find any 80P LVD ultra 160 ones.

Only one I could find in stock was some Seagate Cheetah 36E 18.4GB 4MB / 6ms / 10000RPM / Ultra160 LVD

 

[scsiraid]

There are quite a few 80pin drives on ebay. Note that if you use a non-xSeries drive you will get a warning in ServeRAID manager that you have a non-supported configuration and you will be running a non-tested configuration. U320 drives are backwards compatible with older/slower controllers.

 

[LesterDK]

Thanks for the help, server is back in production, should get replacement drive tuesday.

/Kenneth