1650 RAID 1 lost mirror

Question

I know, it's an old server, but it just keeps going and I can't make myself retire it.   I've got a 1650 with all the latest firmware and such with the on board adaptec RAID and it had a suspected drive failure recently.  It is setup with 2 drives mirrored (RAID1).   The box is running NetWare 5.1sp8.  While the OS was running I pulled the bad drive.  No problems, the system kept on running.  Next I replaced the failed drive with a new drive.  At that point the lights on both the replacement drive and the original still in the machine went orange and the OS basically crashed.   I rebooted the server and received the Reject or Accept new RAID config dialog.  I think I selected reject, but can't be sure.  I went into the RAID BIOS config and the old drive showed up as RAID and the new drive showed up as 'Legacy' and the system started remirroring.   I then rebooted the server again (rebuild was still going on in the background).  The server came up with the C: partition in-tact, but the netware partition was hosed.   Next I pulled both the new drive and the old drive out of the system and put the original, suspected failed drive back in by itself.  The system reported the RAID was critical (as one of the mirrored drives was missing) but the system did boot, and all the data was intact.  So far looks like the failed drive wasn't really dead at all.   My question is, how do you replace a failed drive on a RAID1 set on a 1650?  Like I said, all the firmware is up-to-date, the RAID controller is using 2.8 code and drivers.

DELL-GaryS · Answer

Hi mojoriser,     As you've found out sometimes a drive fails in a manner that messes with communications on the SCSI bus, perhaps causing the wrong drive to go offline. The removal and replacement with the system up is correct, however some troubleshooting is advised.   Regards, Dell-GaryS

mojoriser · Answer

Thanks, GaryS.  Can you be more specific regarding what troubleshooting you recommend?   And, more importantly, what is the process to put a new drive back in to replace the failed one?  I know I *should* be able to hot-swap it, but to minimize risk should I take the box down and put the new drive in?  After doing that and rebooting it's going to ask me about the RAID config.  Should I accept or reject the config?  I'd like to get this system mirrored again.

DELL-GaryS · Answer

Hi mojoriser,     Sorry about the short answer yesterday, your post was already two days old. Some of the trouble shooting would include pulling the controller logs, looking at the properties of the hard drives, reviewing the system event (H/W) logs, and finally running a diagnostics test on the drives. It's been a while for me on netware, generally you use a command line utility for the Adaptec controllers under novell to do the troubleshooting The controller F/W, Driver version and CLI version all should match, we do not advise updating controller F/W with degraded arrays on the system, there should be older versions of the CLI on the file library. CLI app for F/W 2.8  instructions for installation on the page. CLI manual  CLI ref guide   The Cli commands needed are: Open AFA0 Disk list /all Disk show defects (0,(drive ID here),0) repeat for all installed drives you're looking for grown defects (media errors), ignore the primary defects number, it's useless if F/W = > 2.5. Controller details (this will show you your driver and F/W versions) Container list Diagnostics show history /old   If you've installed the new hard drive with power off these commands may be needed: Disk initialize (0,(new hard drive SCSI ID),0) Container set failover 0,(0,(new hard drive ID),0) Controller rescan task list /all     (here you're looking to see a failover task start)   and getting out of the CLI: Close Exit   Regards, Dell-GaryS

PowerEdge HDD/SCSI/RAID

1650 RAID 1 lost mirror

Was this post helpful?