One of my drives reported few read errors (6) through the web interface.

 

After restarting the unRAID server, the count of errors has cleared, but that drive is still marked in red, and does update info such as its temperature, number of read or writes... However the array is online and accessible. I can also read the content from that 'red' drive, and I'm not getting any more read errors (although its stats are not updated in the web ui).

 

My problem is that in addition to being concerned about the well-being of that disk, I also now can not add new drive(s) to the array, as a message states that: "You may not add new disks when there exists missing, wrong, or disabled disk(s)."

 

What do I need to do to clear this status? When I stop the array, I dont get a check party option, but there is an option to 'restore' that states that: "Restore will initialize the stored array configuration; all drives will appear as New, but data disk contents are not affected." - is this what I need to do to clear the error message to enable adding more disks?

 

One of my drives reported few read errors (6) through the web interface.

 

After restarting the unRAID server, the count of errors has cleared, but that drive is still marked in red, and does update info such as its temperature, number of read or writes... However the array is online and accessible. I can also read the content from that 'red' drive, and I'm not getting any more read errors (although its stats are not updated in the web ui).

 

My problem is that in addition to being concerned about the well-being of that disk, I also now can not add new drive(s) to the array, as a message states that: "You may not add new disks when there exists missing, wrong, or disabled disk(s)."

 

What do I need to do to clear this status? When I stop the array, I dont get a check party option, but there is an option to 'restore' that states that: "Restore will initialize the stored array configuration; all drives will appear as New, but data disk contents are not affected." - is this what I need to do to clear the error message to enable adding more disks?

 

A disk marked in "red" has failed.  It needs to be replaced.  The reason you can still get to the contents of the failed drive is that it is being reconstructed from your remaining drives by the unRaid software.  If you didn't have a raid array, you would have lost the contents of the failed drive...... unless another drive fails in your array you will be fine once you replace the failed drive.... do not delay the replacement... currently you are not protected from another failure.

 

DO NOT use the "Restore" option unless you do not want any of the contents on the "red" disk that has failed.  It would  basically forget it existed.

 

So... to clear your "red" error, replace the failed disk.  (It might also be a loose connector, or a bad cable, so check those first after stopping your array and powering down.)

 

To force the unraid server to check the failed drive after you re-seat the connectors you might need to stop the array, un-assign the drive in the drive config page, start the array with it unassigned, stop the array once more, re-assign it to its old slot where it will then think it is a new replacement drive, and then re-start the array.  If it still fails, replace the drive. 

 

Oh yes... if it is an IDE drive, do not use round cables... they cause errors in many installations...

 

After you replace the failed disk and power up again the unRaid software should have a "I'm sure" checkbox below the "Start" button.  Check the box, start the array, and it will first clear the new drive, then format it, then reconstruct the old failed contents onto it.  At that point your  array will be back to where you can add a new drive. 

 

Do one step at a time and you will be fine, but whatever you do, replace the failed drive. (or re-seat its cabling, or replace its bad-cable that is making it look bad)

 

 

Joe L.

Thanks for your help.

 

I followed your instructions, stopped the array, unassigned that bad disk (which is disk1 in my case) and started the array without it.

 

I could still access the content of that disk (obviously through unraid constructing the info through the other drives & the parity disk on the fly).

 

I then stopped the array again, reassigned that drive to the same slot (hoping it may be physically okay and it was only a connection or power problem with it), and started the array.

 

It kept on 'starting' with nothing happening on the web-ui.

Telneting into the tower and tailing the log file, it is full of errors: "md: bug in file unraid.c line 201". The last message in the log is "reiserfs: checking transaction log (device md(9,4)) ..." and it stalls there.

 

Rebooting the machine gets to the same state, so I cant unassign that drive and start the array without it again to assign it to another physical drive.

 

It could be that some check is running that will finish or timeout and indicate the drive as failed again, at which time I'll regain access to the web-UI through which I could stop the array, and of course its also possible its in some infinite loop...

 

Should I wait, or is there a command line for stopping the array so I can then unassign that disk?

BTW - the web-UI now shows both the disk1 (that was previosuly 'red') and the parity as 'orange' and all other six drives are green.

 

Thanks in advance,

Shahar

After few more hours: array still doesnt get started, nor do I get a way to stop it through the web-ui and unmount that drive.

 

The log has been updated with several "smbd/service.c:make_connection(800)" that result in "couldn't find service all"

 

And no progress since.

Waiting for suggestions on what to do next. My guess is some command line to manualy stop the array and unmount that drive?

You can try unmenu to see what's going on with the array.  Check out the thread http://lime-technology.com/forum/index.php?topic=153.0 It's simple to add to your flash drive and comes in handy.

After few more hours: array still doesnt get started, nor do I get a way to stop it through the web-ui and unmount that drive.

 

The log has been updated with several "smbd/service.c:make_connection(800)" that result in "couldn't find service all"

 

And no progress since.

Waiting for suggestions on what to do next. My guess is some command line to manually stop the array and unmount that drive?

On most versions of unRaid there exists a "stop" command that is available once you log in via telnet.

Simply type stop

 

If it does not exist, you can perform the same steps with this series of commands:

samba stop

sync

umount /mnt/disk1

umount /mnt/disk2

umount /mnt/disk3

umount /mnt/disk4

umount /mnt/disk5

umount /mnt/disk6

umount /mnt/disk7

umount /mnt/disk8

umount /mnt/disk9

umount /mnt/disk10

umount /mnt/disk11

umount /mnt/disk12

umount /mnt/disk13

mdcmd stop

 

It will stop the array.

 

There is also a 'poweroff' command to shut down the server.

 

Before stopping the array typing mount will show you the drives currently mounted.  Stopping the array will un-mount all the data drives and stop samba file sharing.

 

 

Joe L.

Thanks again for your response, but alas still no luck. I'm without my data for too long now. Its no longer fun 

 

When the server boots, I immediately get to the same state with the web-ui saying its starting, and staying there forever.

 

When telneting in and running "stop" I get that no disks were mounted... but the web-ui remains on 'starting', and still doesn't let me stop the array / remove that disk.

Tailing syslog shows mdcmd (9) stop, (11) stop, (18) stop and (22) stop - all of which get a "md: 5 devices still in use.".

 

Btw - possibly its important: on the console screen, right after boot and before the tower login prompt I get:

"mount: wrong fs type, bad option, bad superblock on /dev/md1, missing codepage or other error" and the same message for /dev/md3. I also get the helpful suggestion that "in some cases useful info is found in syslog - try dmesg tail".

 

What's going so wrong?

More info: using the unraid telnet menu, I can see that out of the 8 connected disks, disk0 (parity) and disk1 both show as disk_invalid.

Doesn't sound good.

Suggestions please?

Hi ShaharT,

Please send email to [email protected] and I'll try to help you get things sorted out.

Hi, done that few hours ago, I'll resend now in case your spam blocked it.

I sure hope there's a way to get this fixed.

Thanks in advance,

Shahar

To close the loop on this one:

Tom very kindly reviewed the log files and super.dat, and found out that for some odd reason super.dat got screwed up. He sent me a fixed one which got me back to the starting point (disk 1 down). I then replaced that disk, had the content reconstructed, and carried on to add 3 more disks (now with 11 disks) as well as upgraded to 4.0b9 - all working fine.

Many thanks for the outstanding support.

Shahar