I apologize if this gets posted in multiple areas.
I have the following message spooling to the console of my K200. This machine is no longer under HP hardware maintenance.
Data redundancy has been lost in disk array on /dev/dsk/C4t0d0. Disk in slot B4 has failed.
Active hot spare warning on /dev/rdsk/C4t0d0. Unallocated space is too small for active hot spare.
The rebuild is currently at 62%. The array has 7 18.2 GB disks.
Once I get the new disk from my vendor, how should I proceed to replace it? Is it really hot-swapable? Is it actually hands-off? I am new to this particular crisis, so I appreciate your responses very much.
It looks like your 12H is probably running RAID5, hence the rebuild message once a disk failed. What you should do is wait till the rebuild is completed before replacing the disk at slot B4. And yes (if your replacement disk is new) this is a hot swap operation. If the replacement disk is "used" (ie from another 12H) then you have to use the front panel of the 12H to "include" this disk into the array. To check is the rebuilt has completed, run this command .. # arraydsp -i ==> take note of the serial number of the array # arraydsp -a ==> Look for the line "Overall State of Array". Since you have time .. can you post the output of .. arraydsp -a .. this would help in verifying what status and the configuration of your array. You may want to consider activating the hot spare space. Post the output and we'll see what we can do.
Absolutely, simply remove disk B4, pop in the new drive and walk away; however, this only applies if 'Auto Include' has been enabled which is the norm. Otherwise, you have to use the front panel menu Disk Changes -> Include Disk -> B4 to start the operation after the module has been replaced.
In most cases, it's just pop the disk in and walk away. DO NOT REPLACE ANOTHER DISK until the rebuiild is finished.
The specifications I remember from that unit(we never bought one, just shopped it) says it has hot swapable drives.
They have a little tab on the bottom, you simply pull and the dead drive is out. Insert new drive. You get an lbolt, you are done.
The reall issue is making sure you get the RIGHT drive. Be sure.
SEP
Steven E Protter Owner of ISN Corporation http://isnamerica.com http://hpuxconsulting.com Sponsor: http://hpux.ws Twitter: http://twitter.com/hpuxlinux Founder http://newdatacloud.com
From your attachment we can see that the rebuild is now 89% complete. You should wait for the rebuild to complete, then simply replace the drive. At that point, we will find out whether "auto include" is enabled or not. If it is, you're all set - if not, you'll need to include it manually via the panel.
The difference is, if it's a new disk you can simply plug it and the array will "do its thing automatically". Again make sure the rebuild is completed before doing this operation. Id the disk is used (for example if it came from another 12H), then you need to do this .. Press Menu, then "+" until you see "Disk Changes" and then press Enter. Somewhere in there there should be an option that says "Include Disk", choose that one to perform the include operation of this disk into the array. See this my reply in this thread .. http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x24199607df6ed711abdc0090277a778c,00.html
From your arraydsp -a output, the line .. Used by non-included disks = 17366 MB which represent the size of a single disk confirms that one of the disk is not functioning. In a normal 12H this line should say 0 MB. After you plug in the new disk, run arraydsp -a again and you should see that change plus the active hot spare space should come back (since hot spare is enabled) .. form the line .. Active Hot Spare Desired = ENABLED Later on you may want to followup on this thread and re-post arraydsp -a just to make sure. Just a suggestion ..
One little note: when rebuild will complete you can try to remove and insert this disk back which is in failed state now. Failure condition will be cleared, and balancing will start. If disk is really fauly, balancing will be cancelled and disk will be faulted again. Then surely replace this disk. Why? There were cases when autoraid was failing disk, but disk was not really bad. Only logs will show it (logprint -t All -v) Eugeny
The rebuild has completed and here is an updated arraydsp output.
I made a mistake earlier, there are 8 18.2 disks in the array.
My assumptions:
Used as Active Hot spare = 0 b/c the spare disk took over for B4 (failed)
I have an additional question, though. My machine rebooted when the disk failed. Is that related to the prior "Unallocated space is too small for active hot spare."? I thought that with hot spare enabled the array manager would hold back a drive as the hot spare. I guess I don't know why the reboot. I would have thought the hot spare would have kicked in w/o missing a beat.
Again, thank you for your response and I will assign points once things slow a bit and the machine is recovered.
The 'Active Hot Spare' of an AutoRAID is more complicated than that. The spare does not do a one for one (usually) replacement and is used to increase the amount of RAID 1/0 space. Once there was a failure there was not enough unallocated space to allow for the additional failure of the largest drive remaining.
The machine should not have rebooted or hung. In fact, the failure should have been all but invisible to the OS other than the array monitoring daemon.
I would like to see the output of arraydsp -a after the replacement of the disk in B4. It's basically still showing the same information except for the status of the rebuild because it has completed its rebuild process. Even without a hotspare on a 12H, there should be enough redundancy to take care of ONE faulty disk. The 12H will allocate that amount of required space in "used for redundancy" automatically. Hence failure of one disk should not cause your system to reboot. Check the system logs, GSP logs, etc for clues. BTW hot spare allocating is not tie to one particular disk. In a 12H you can't tell which disk is used as hotspare. The internal machanism of this array may allocated spaces from all disks to add up to the size of the largest disk in the group representing the hotspare.
At this point all you need to do is plug in the new disk once you got it and check the status of the array periodically to make sure everything is back to normal.
