Foreign Configuration

So, i imported the config and in the BIOS it reported failed. I took the drive out and reinserted it hot. The light on the front of the drive is green and in the BIOS it reports foreign config. I am thinking about clearing the foreign config and i have reinserted the original failed drive. Maybe that will return the server back to 1 failed drive and from there i can replace drive 3? I need someones help :)

This second array houses our exchange DB so this is a very hot issue for me...

I'm not sure I follow ... what is the current status of each drive and each array? How many arrays do you have and which drives are members of which?

What is the proper procedure to add the missing drive and deal with the foreign configuration?

Actually, this is backwards. You need to deal with the foreign configuration first, THEN "add" the missing drive.

In the CTRL-R utility, on the VD MGMT screen, highlight the controller, F2, Foreign, Import. If the import is successful and the OS is functional and the data intact, then you can deal with the missing disk; if it is in a Ready state, then just assign it as a hot-spare to start the rebuild.

So the server has 8 total drives. Drive 0, 1 is 1 array (contains the OS server 2012). Drive 2-7 are 1 array (labeled data in the OS and houses hyper-v VHDX's). The status in CTRL-R of drive 0 and 1 are "online". Status of drives 2 and 3 in CTRL-R, 2=foreign and 3=failed. Drives 4-7 are status online.

The import did fail and when the import failed, drive 2's status was failed. I then removed drive 2 and then reinserted it. This brought the status back to foreign for drive 2. This is where we currently are.

Array (VD0) 0 status is optimal and array (VD1) 1 is offline

Anyone know how to save the array?

If the import fails, there are limited options. This only applies because you said that disk 3 was already failed before anything else happened.

I am not sure if disk 3 failed first or not. I only noticed them in the failed state at the same time

- Best option is to import the foreign configuration. If you can't import it, then things get sketchy. - If disk 2 shows "failed", you can force it online.

So, i should import and if that import fails, force drive 2 online and then reboot and maybe this will save the array?

- You could perform a "retag". A retag is when you delete the array and create it anew WITHOUT initializing it.

Retag does not delete the data on the array?

What is the PERC 6 firmware version (CTRL MGMT screen, package version)?

If the import fails, there are limited options.

This only applies because you said that disk 3 was already failed before anything else happened.

- Best option is to import the foreign configuration. If you can't import it, then things get sketchy.

- If disk 2 shows "failed", you can force it online.

- You could perform a "retag". A retag is when you delete the array and create it anew WITHOUT initializing it.

What is the PERC 6 firmware version (CTRL MGMT screen, package version)?

The import failed and i forced online. When rebooting the server, disk 2 went to flashing amber. Here is the controller version:

BTW, there is a hotspare in the server and i have access to 2 PERC 6/I controllers that might have more current firmware (if these things help).

I am not sure if disk 3 failed first or not. I only noticed them in the failed state at the same time

Sorry, this is what lead me to say that:

i noticed physical drive 3 blinking amber. When i replaced it with same capacity drive, i must have inadvertently released physical drive 2 loose

So, i should import and if that import fails, force drive 2 online and then reboot and maybe this will save the array?

Maybe. It can. It is kind of a last-ditch effort to put members of the array back in the positions they need to be in. Don't do this until you have more instructions, and don't do this if you don't have a backup and may need to consider professional data recovery.

Retag does not delete the data on the array?

Not when done correctly.

Updating the firmware "could" help, but I would not hold my breath. It may be a decent "last-ditch" effort to be able to import the foreign configuration properly.

carnold5 I am not sure if disk 3 failed first or not. I only noticed them in the failed state at the same time Sorry, this is what lead me to say that:

To be honest, now so many things have been done and this all took place on 7.2, i cant remember what took place first.

Updating the firmware "could" help, but I would not hold my breath. It may be a decent "last-ditch" effort to be able to import the foreign configuration properly.

I updated the firmware and then imported the foreign config. This took drive 2 into the failed state and i then forced online. This took array 1 to degraded. Boot the server and still missing drive D:.

In windows (i can manage the server using server manager), if i launch computer management, the OS wants to initialize the drive (i did not initialize the drive). It shows up as unknown.

One thing that is a little weird, in CTRL-R, drive 2 either shows as foreign config (PD MGMT) or when imported it shows as failed (VD MGMT). drive 3 all the time shows as failed in PD MGMT and missing in VD MGMT. Shouldn't drive 3 also shows as failed and not missing?

One thing that is a little weird, in CTRL-R, drive 2 either shows as foreign config (PD MGMT) or when imported it shows as failed (VD MGMT). drive 3 all the time shows as failed in PD MGMT and missing in VD MGMT. Shouldn't drive 3 also shows as failed and not missing?

Not necessarily. If disk 3 is showing missing, it may very well be bad - drives don't have to be bad to fail, but if it is bad, then it may simply not be communicating with the controller. Also, even if you had two foreign disks, importing would be successful but would leave 1 of the disks in a failed state. This is to preserve the array in the event that the foreign disk contains corrupt data - the controller will not start a quick init of the array, further corrupting the data.

A simple data recovery utility might be able to show you if the data is still intact on the data volume, and may be able to simply repair the damaged partition tables, etc., but if it doesn't find any data or cannot repair it, then data recovery or backups are the next step.

Also, even if you had two foreign disks, importing would be successful but would leave 1 of the disks in a failed state. This is to preserve the array in the event that the foreign disk contains corrupt data

Not sure where to go from here? 1 thing: with drive 2 in the foreign config state, can i replace drive 3 and it rebuild then deal with the foreign config?

I havent tried to delete the array and then recreate without initializing.

with drive 2 in the foreign config state, can i replace drive 3 and it rebuild then deal with the foreign config

No, remember I commented in the very beginning that this is approaching the problem backwards. Again, this is assuming that disk 3 failed first - and I know you said now that you aren't sure, but it makes all the difference.

RAID 5 allows for ONE disk to fail before the array goes offline. If one disk goes offline, you can rebuild (or replace) it. If two disks go offline, you can force one online (you don't want to force both for reasons mentioned above), BUT if one disk failed before the other, and you force online the disk with out-of-date data, you will corrupt the data, possibly resulting in a completely unusable volume.

So, 1) disk 3 seems to be non-functional and you can't replace it because you can't rebuild a drive in an offline array (offline means too many disks are missing and data is incomplete, unreadable, and inaccessible), and 2) even if disk 3 was functional, if it was the first to fail, it would likely corrupt your data if you forced it online.

So, with disk 3 non-functional anyway, you really have no choice but to work with disk 2's configuration and hope you can get the array online with that.

carnold5 with drive 2 in the foreign config state, can i replace drive 3 and it rebuild then deal with the foreign config No, remember I commented in the very beginning that this is approaching the problem backwards. Again, this is assuming that disk 3 failed first - and I know you said now that you aren't sure, but it makes all the difference.

Yea, i know, i just want to exhaust all the possibilities before i give up the ghost...

RAID 5 allows for ONE disk to fail before the array goes offline. If one disk goes offline, you can rebuild (or replace) it. If two disks go offline, you can force one online (you don't want to force both for reasons mentioned above), BUT if one disk failed before the other, and you force online the disk with out-of-date data, you will corrupt the data, possibly resulting in a completely unusable volume. So, 1) disk 3 seems to be non-functional and you can't replace it because you can't rebuild a drive in an offline array (offline means too many disks are missing and data is incomplete, unreadable, and inaccessible), and 2) even if disk 3 was functional, if it was the first to fail, it would likely corrupt your data if you forced it online. So, with disk 3 non-functional anyway, you really have no choice but to work with disk 2's configuration and hope you can get the array online with that.

The only thing left to do is delete the array and rebuild without initializing. You said earlier about doing it correctly. What is the correct way to delete an array and then recreate the array (especially with non-functional drive 3)?

We cant do anything with the hotspare that is in the array (drive 7)?

When you do the retag, you delete the VD, then create it again - but the number and positions of the disks must be the same, as does the stripe size. You need to leave the Advanced box UNCHECKED.

Since drive 3 is non-functional, there has to be a drive to take its place ... you will need to put an empty drive in its place, and as SOON as you create the new array, you need to force drive 3 offline, because as soon as the array has been created, it will begin reconciling data across the drives and if it encounters drives with irrelevant data, it will start making decisions, which will ultimately end in corruption.

The hot-spare is a non-participating member of the array - it has NO data on it that is relevant to the array - it is only standing ready to start rebuilding in the event a disk fails.

If you need to consider professional data recovery, I would do so before performing a retag.