Help rebuild SATA Mirrored Pair

[Dead Dead Money]

I have an Asus P4G8X motherboard with Silicon Image 3112a Controller
and I installed a set of new SATA hard drives a few weeks ago as a
mirrored pair - everything worked fine. (Only drives in the system.)

Yesterday when booting up I got an error, "incomplete raid set" then
before I could react the system booted into Windows XP Pro.

I restarted the system and hit ctrl S to get to the RAID Configuration
Utility, selected the option to rebuild the mirrored set. Selected the
option for offlinerebuild and entered Y to start. Nothing appeared to
happen but I let it go for a few hours but nothing happened.

Checked around and found that there should be some indication of
percentage that was complete via a progress meter but it never even
appeared.

I updated the RAID utility to what I think is the latest version,
dated 2003 and tried again - no difference nothing happened.

I updated the BIOS to the latest version I could find - Rev 1006 and
tried again - still nothing.

I tried the option for onlinerebuild - still nothing.

The utility shows Drive 0 Current and Drive 1 Rebuild

The computer will boot into Windows and appears to run normally with
the exception of the drive.

Any help would be greatly appreciated!

TIA

 

[Arno Wagner]

I have an Asus P4G8X motherboard with Silicon Image 3112a Controller
and I installed a set of new SATA hard drives a few weeks ago as a
mirrored pair - everything worked fine. (Only drives in the system.)

Yesterday when booting up I got an error, "incomplete raid set" then
before I could react the system booted into Windows XP Pro.

I restarted the system and hit ctrl S to get to the RAID Configuration
Utility, selected the option to rebuild the mirrored set. Selected the
option for offlinerebuild and entered Y to start. Nothing appeared to
happen but I let it go for a few hours but nothing happened.

Checked around and found that there should be some indication of
percentage that was complete via a progress meter but it never even
appeared.

I updated the RAID utility to what I think is the latest version,
dated 2003 and tried again - no difference nothing happened.

I updated the BIOS to the latest version I could find - Rev 1006 and
tried again - still nothing.

I tried the option for onlinerebuild - still nothing.

The utility shows Drive 0 Current and Drive 1 Rebuild

The computer will boot into Windows and appears to run normally with
the exception of the drive.

Any help would be greatly appreciated!

Does your RAID BIOS detect both disks? If not, you need to replace
the broken disk before rebuilding. Any disk at least the same
size (in sectors, one with the same number og GBs meu be a bit smalle)
should work.

Arno

 

[Dead Dead Money]

Does your RAID BIOS detect both disks? If not, you need to replace
the broken disk before rebuilding. Any disk at least the same
size (in sectors, one with the same number og GBs meu be a bit smalle)
should work.

Arno

Hi Arno, thanks for the reply.

Yes both drives show up when booting and in the RAID Utility (crtl S
during bootup)

Interestingly while searching the Asus site I found someone with a
differently model motherboard who, seemingly, has the same problem -
"incomplete raid set" and when attempting to rebuild, no progress
meter. (no reply to either of us from Asus LOL)

Does the "incomplete raid set" typically mean that the system is not
seeing one of the drives? (I'm a total noob on RAID but this is what
I suspect your question implied...)

Thanks again for your reply!

 

[Arno Wagner]

Does your RAID BIOS detect both disks? If not, you need to replace
the broken disk before rebuilding. Any disk at least the same
size (in sectors, one with the same number og GBs meu be a bit smalle)
should work.

Arno

Hi Arno, thanks for the reply.[/QUOTE]

Yes both drives show up when booting and in the RAID Utility (crtl S
during bootup)

Interestingly while searching the Asus site I found someone with a
differently model motherboard who, seemingly, has the same problem -
"incomplete raid set" and when attempting to rebuild, no progress
meter. (no reply to either of us from Asus LOL)

Does the "incomplete raid set" typically mean that the system is not
seeing one of the drives? (I'm a total noob on RAID but this is what
I suspect your question implied...)

It implies that one drive is missing from the RAID set. The reason
can be that it has been kicked from the set due to it having misbehaved
(errors, long answering delay) or it can have failed to be detected.
Any decent RAID cpntroller will give you a report on why the drive
is missing (although I suspect that these ElCheapo controllers do
not).

So the next thing you should do is to find out why the drive was
removed from the RAID. For this you should first identify which drive
forms the (now one disk) RAID1 array and which one is the kicked
one. Then you should get the SMART attributes from the kicked one.

If it is fine, then you should be able to add it to the existing
RAID and rebuild. If it is not fine, you need to replace it
before rebuilding.

BTW, you are running without redundancy at the moment, I hope your
backups are current...

Arno

 

[Folkert Rienstra]

Does your RAID BIOS detect both disks? If not, you need to replace the
broken disk before rebuilding. Any disk at least the same size (in sectors,

one with the same number og GBs meu be a bit smalle) should work.

The babblebot seems to be finally running out of braincells.

 

[Folkert Rienstra]

Hi Arno, thanks for the reply.

Yes both drives show up when booting and in the RAID Utility (crtl S
during bootup)

Interestingly while searching the Asus site I found someone with a
differently model motherboard who, seemingly, has the same problem -
"incomplete raid set" and when attempting to rebuild, no progress
meter. (no reply to either of us from Asus LOL)

Does the "incomplete raid set" typically mean that the system is not
seeing one of the drives?

You just answered that question right above.
Please do not encourage the braindamaged "og GBs meu be a bit smalle" babblebot.
On a good day it can't even write it's own name.

 

[Dead Dead Money]

It implies that one drive is missing from the RAID set. The reason
can be that it has been kicked from the set due to it having misbehaved
(errors, long answering delay) or it can have failed to be detected.
Any decent RAID cpntroller will give you a report on why the drive
is missing (although I suspect that these ElCheapo controllers do
not).

So the next thing you should do is to find out why the drive was
removed from the RAID. For this you should first identify which drive
forms the (now one disk) RAID1 array and which one is the kicked
one. Then you should get the SMART attributes from the kicked one.

If it is fine, then you should be able to add it to the existing
RAID and rebuild. If it is not fine, you need to replace it
before rebuilding.

BTW, you are running without redundancy at the moment, I hope your
backups are current...

Arno

Hi Arno,

Based on your reply about the drive missing I tried swapping out the
data cable on the "missing" drive and it came back up!

I guess it is time to search out some QUALITY SATA cables...

One last question, hopefully. When the system rebooted without errors
after I replaced the cable I checked the RAID configuration utility
and it showed the drive back in action.

I thought I would go ahead and try to rebuild the array just to make
sure the drives were really synced up however it wouldn't let me do
it.

My question is would the drives "automagically" resync themselves? I
didn't purposely change any data files during the time the one drive
was dead however I figured some system files (swap file, browser
history etc) would have changed just during the times I rebooted and
surfed around looking for possible solutions.

Thanks for all your help, it is greatly appreciated!

 

[Arno Wagner]

On 10 Jul 2007 17:37:27 GMT, Arno Wagner <[email protected]> wrote:

Hi Arno,

Based on your reply about the drive missing I tried swapping out the
data cable on the "missing" drive and it came back up!

Cool! That also explains why the drive was removed from the
array: Too many interface errors. The initial drive detection
is less critical and could still have worked with a bad cable.

I guess it is time to search out some QUALITY SATA cables...

I have had problems with some too.

One last question, hopefully. When the system rebooted without errors
after I replaced the cable I checked the RAID configuration utility
and it showed the drive back in action.

I thought I would go ahead and try to rebuild the array just to make
sure the drives were really synced up however it wouldn't let me do
it.

My question is would the drives "automagically" resync themselves?

Usually not.

I
didn't purposely change any data files during the time the one drive
was dead however I figured some system files (swap file, browser
history etc) would have changed just during the times I rebooted and
surfed around looking for possible solutions.

That is not how it works. The RAID controller does not know or understand
files. In fact it does not know or undertsand the filesystem in the first
place. But I seem to remember you said something about a "background
reconstruction" earlier? That would be done while the array was in use.

To be sure, check the array status with the monitoring/management
software. You should be able to get that from the manufacturers
(either controller or mainboard) website.

Arno

 

[Dead Dead Money]

Cool! That also explains why the drive was removed from the
array: Too many interface errors. The initial drive detection
is less critical and could still have worked with a bad cable.


I have had problems with some too.




Usually not.


That is not how it works. The RAID controller does not know or understand
files. In fact it does not know or undertsand the filesystem in the first
place. But I seem to remember you said something about a "background
reconstruction" earlier? That would be done while the array was in use.

To be sure, check the array status with the monitoring/management
software. You should be able to get that from the manufacturers
(either controller or mainboard) website.

Arno

Hi Arno

Sorry it took me so long to get back to you - I had some other
problems LOL!!!

I checked the Silicon Image site for a monitoring/management utility
but they seem refer you back to the motherboard manufacture.

Checking on the Asus I wasn't able to find anything there that would
let me check the drives - only add or delete members or rebuild
mirrored set - however that option doesn't do anything!

Is there some generic utility that would let me compare the drives or
are they proprietary for the controller?

If not could I just delete the array then rebuild it?

Thanks again for putting up with all this!

Hope you have a good weekend!

 

[Arno Wagner]

Hi Arno

Sorry it took me so long to get back to you - I had some other
problems LOL!!!

I checked the Silicon Image site for a monitoring/management utility
but they seem refer you back to the motherboard manufacture.

Checking on the Asus I wasn't able to find anything there that would
let me check the drives - only add or delete members or rebuild
mirrored set - however that option doesn't do anything!

Is there some generic utility that would let me compare the drives or
are they proprietary for the controller?

Well, under Windows I fear all this is proprietary. Under Linux
there has been some reverse-enginnering and building of
non-proprietary tools, but it seems these cheap RAID controllers
are generally a sad business...

If not could I just delete the array then rebuild it?

If you do not mind looing your data, that should work. It will
still not give you any morniroting while the system is running.

Thanks again for putting up with all this!

No problem.

Arno