I have moved my 9 drive setup to a norco 4220 case with an MSI P6N Diamond motherboard, and an SAS LP-MV8 card. 8 of the drives are now connected to the SAS card and the 9th directly to the motherboard SATA 1 connector.  I have correctly assigned all the drives to their correct position under the "Devices" tab; however, now 2 of the drives appear as Unformatted.

 

Please help! I am positive that all drives are assigned properly and have rebooted the system. On the first (initial) boot, 3 drives appeared as unformatted, now 2. I do not want to keep rebooting, but am not sure what else I can do.

 

I am leaving the system as is till someone suggests something to correct the issue. Syslog attached.

 

Thanks.

syslog-2010-06-22.txt.zip

Your troubled disks are md6 and md7. Later in the syslog looks like there's additional drive status errors as well as trouble reading certain sectors.

 

 

Jun 21 23:29:43 Tower kernel: md: import disk6: [8,48] (sdd) ST31500341AS    9VS0TAJ9 offset: 63 size: 1465138552

Jun 21 23:29:43 Tower kernel: md: import disk7: [8,64] (sde) ST3500630AS      9QG1FGFM offset: 63 size: 488386552

Jun 21 23:29:44 Tower kernel: md6: running, size: 1465138552 blocks

Jun 21 23:29:44 Tower kernel: md7: running, size: 488386552 blocks

 

Jun 21 23:29:44 Tower logger: mount: wrong fs type, bad option, bad superblock on /dev/md6,

Jun 21 23:29:44 Tower logger:        missing codepage or helper program, or other error

Jun 21 23:29:44 Tower logger:        In some cases useful info is found in syslog - try

Jun 21 23:29:44 Tower logger:        dmesg | tail  or so

Jun 21 23:29:44 Tower logger:

Jun 21 23:29:44 Tower emhttp: _shcmd: shcmd (22): exit status: 32

Jun 21 23:29:44 Tower emhttp: disk6 mount error: 32

 

Jun 21 23:29:44 Tower emhttp: shcmd (23): rmdir /mnt/disk6

Jun 21 23:29:44 Tower logger: mount: wrong fs type, bad option, bad superblock on /dev/md7,

Jun 21 23:29:44 Tower logger:        missing codepage or helper program, or other error

Jun 21 23:29:44 Tower logger:        In some cases useful info is found in syslog - try

Jun 21 23:29:44 Tower logger:        dmesg | tail  or so

Jun 21 23:29:44 Tower logger:

Jun 21 23:29:44 Tower emhttp: _shcmd: shcmd (23): exit status: 32

Jun 21 23:29:44 Tower emhttp: disk7 mount error: 32

 

Jun 21 23:29:44 Tower kernel: ata9: translated ATA stat/err 0x01/04 to SCSI SK/ASC/ASCQ 0xb/00/00

Jun 21 23:29:44 Tower kernel: ata9: status=0x01 { Error }

Jun 21 23:29:44 Tower kernel: ata9: error=0x04 { DriveStatusError }

Jun 21 23:29:44 Tower kernel: ata9: translated ATA stat/err 0x01/04 to SCSI SK/ASC/ASCQ 0xb/00/00

Jun 21 23:29:44 Tower kernel: ata9: status=0x01 { Error }

Jun 21 23:29:44 Tower kernel: ata9: error=0x04 { DriveStatusError }

Jun 21 23:29:44 Tower kernel: ata9: status=0x01 { Error }

Jun 21 23:29:44 Tower kernel: ata9: error=0x04 { DriveStatusError }

 

Jun 21 23:29:44 Tower kernel: sd 7:0:2:0: [sde] Result: hostbyte=0x00 driverbyte=0x08

Jun 21 23:29:44 Tower kernel: sd 7:0:2:0: [sde] Sense Key : 0xb [current] [descriptor]

Jun 21 23:29:44 Tower kernel: Descriptor sense data with sense descriptors (in hex):

Jun 21 23:29:44 Tower kernel:        72 0b 00 00 00 00 00 0c 00 0a 80 00 00 00 00 00

Jun 21 23:29:44 Tower kernel:        00 00 00 be

Jun 21 23:29:44 Tower kernel: sd 7:0:2:0: [sde] ASC=0x0 ASCQ=0x0

Jun 21 23:29:44 Tower kernel: sd 7:0:2:0: [sde] CDB: cdb[0]=0x28: 28 00 00 01 00 cf 00 00 08 00

Jun 21 23:29:44 Tower kernel: end_request: I/O error, dev sde, sector 65743

Jun 21 23:29:44 Tower kernel: md: disk7 read error

Well, first, DO NOT click 'Format'.

 

The system log is showing massive amount of error output for disk6 and disk7 (maybe others too).

 

This is probably a backplane issue, cable issue, or bad controller, or possibly an incompatibility between the controller and your motherboard (though probably not since some of the drive are fine).  The card is a PCI-E x4 card so it has to be plugged into a x4, x8, or x16 PCI-E slot.

 

Time to start troubleshooting hardware.

 

Some guidelines:

1. Look at your system log and get an idea of what the error messages look like.  For instance, all these types of lines are not good:

Jun 21 23:29:44 Tower kernel: /usr/src/sas/trunk/mvsas_tgt/mv_sas.c 2071:Port 2 irq sts = 0x1000000

 

2. Unassign your parity drive.

3. Remove disk6 and disk7.

4. Boot up & from console or telnet type:

initconfig

This will clear the configuration.  (Used to be a button that did this, but Joe made me take it off, but I'm going to put it back next release.)

5. Now refresh the Main page and Start array and then look in your system log... do you see any errors?  If not, it's isolated to these two drives.

6. If no errors in step 5, you need to move those drives to different slots, or move known working drives into those slots... what you are trying to do is isolate which slots cause the problem.

7. You can swap cables around seeing if the problem follows the slot or the cable.

 

During this process, take a good look at the text beside the 'Start' button.  Do not proceed if says anything like "This will clear drives first".  Do not ever click a Format button if one appears.

 

You get the idea?

 

During this process, as long as you don't write to any array drive, when it comes time to re-install Parity there is a way to add it into the array without having to do a parity-sync:

 

1. Install parity & assign via Devices page.

2. Start server & look on Main page, array should not Stopped.

3. From telnet or console type these commands:

initconfig

4. Refresh the Main page, array status should show "all new drives".

5. From telnet or console type this:

mdcmd set invalidslot 99

6. Click 'Start' button on Main - array should show up with all 'green' drive status.

7. Run a parity-check for the fun of it.

 

Thank you for the reply!

 

 

 

1. Look at your system log and get an idea of what the error messages look like.  For instance, all these types of lines are not good:

Jun 21 23:29:44 Tower kernel: /usr/src/sas/trunk/mvsas_tgt/mv_sas.c 2071:Port 2 irq sts = 0x1000000

 

2. Unassign your parity drive.

3. Remove disk6 and disk7.

4. Boot up & from console or telnet type:

initconfig

This will clear the configuration.  (Used to be a button that did this, but Joe made me take it off, but I'm going to put it back next release.)

5. Now refresh the Main page and Start array and then look in your system log... do you see any errors?  If not, it's isolated to these two drives.

6. If no errors in step 5, you need to move those drives to different slots, or move known working drives into those slots... what you are trying to do is isolate which slots cause the problem.

7. You can swap cables around seeing if the problem follows the slot or the cable.

 

During this process, take a good look at the text beside the 'Start' button.  Do not proceed if says anything like "This will clear drives first".  Do not ever click a Format button if one appears.

 

You get the idea?

 

During this process, as long as you don't write to any array drive, when it comes time to re-install Parity there is a way to add it into the array without having to do a parity-sync:

 

1. Install parity & assign via Devices page.

2. Start server & look on Main page, array should not Stopped.

3. From telnet or console type these commands:

initconfig

4. Refresh the Main page, array status should show "all new drives".

5. From telnet or console type this:

mdcmd set invalidslot 99

6. Click 'Start' button on Main - array should show up with all 'green' drive status.

7. Run a parity-check for the fun of it.

 

 

I just want to make sure that I am doing everything correct. Later on tonight, I'll check all the connections.

 

When following your outline, do I reboot at any time? When do I need to generate another Syslog in step 5?

 

Thank you again, I'll post as soon as I follow these steps.

 

Lev

 

(Used to be a "Restore" button that did this, but Joe made me take it off, but I'm going to put it back next release.)

I never asked it be removed.  I've asked for years that it be re-named.    We've just not been able to settle on a different name.

 

If you put a "button" back, at least re-name it to

"Initialize Configuration"

and change the affiliated description to indicate it "Sets a new initial configuration and immediately invalidates any prior parity calculations based on any previous configuration."  State the the new configuration will include only the currently assigned and working drives. 

 

Also, update your on-line manual in the wiki to reflect whatever you finally name it.  Otherwise, users will complain when the manual does not match what they are seeing on the screen.

 

Joe L.

(Used to be a "Restore" button that did this, but Joe made me take it off, but I'm going to put it back next release.)

I never asked it be removed.  I've asked for years that it be re-named.    We've just not been able to settle on a different name.

 

If you put a "button" back, at least re-name it to

"Initialize Configuration"

and change the affiliated description to indicate it "Sets a new initial configuration and immediately invalidates any prior parity calculations based on any previous configuration."  State the the new configuration will include only the currently assigned and working drives. 

 

Also, update your on-line manual in the wiki to reflect whatever you finally name it.  Otherwise, users will complain when the manual does not match what they are seeing on the screen.

 

Joe L.

 

I agree with Joe 

OK. I've done:

 

2. Unassign your parity drive.

3. Remove disk6 and disk7.

4. Boot up & from console or telnet type:

initconfig

 

Also, thinking that maybe the connections are possibly bad, I've moved the drives around a bit - a little shuffling.  I am given the option to "Start will bring the array on-line, start Data-Rebuild, and then expand the file system (if possible)."

 

The Disk 7 - has a [glow=blue,2,300]blue[/glow] ball next to it.

Parity Disk - [glow=orange,2,300]Orange[/glow] ball.

 

What do I do next?

 

Latest syslog attached.

 

Thanks.

syslog-2010-06-22b.txt.zip

OK. I've done:

 

2. Unassign your parity drive.

3. Remove disk6 and disk7.

4. Boot up & from console or telnet type:

initconfig

 

Also, thinking that maybe the connections are possibly bad, I've moved the drives around a bit - a little shuffling.  I am given the option to "Start will bring the array on-line, start Data-Rebuild, and then expand the file system (if possible)."

 

The Disk 7 - has a [glow=blue,2,300]blue[/glow] ball next to it.

Parity Disk - [glow=orange,2,300]Orange[/glow] ball.

 

What do I do next?

 

Latest syslog attached.

 

Thanks.

 

If you unassigned parity in the Devices page, it's not possible for it to have an orange ball next to it on the Main page.

You are absolutelly correct. While shuffling the drives just now I did reassign it. Sorry. What do I do next? I have not changed anything else since the last syslog.

Think about it this way.

 

Let suppose you have parity unassigned.  You can leave the physical disk in the system.

 

Now plug your data drives into any of the other slots in your chassis.

 

1. Power up server and let it boot.  From the console, log in and type 'initconfig'.  Now fire up the webGui and look at the Main page.  It should show "all new drives" (blue dots next to them all).

 

2. Click Start.  unRAID will record which disks are in which slots.  Then mount all the disks.

Did any come up unformatted?  Look in the system log; any errors?

 

If yes, then you have h/w problems in those slots.

 

3. Say you now want to rearrange which disks/slots are used.  Power down.  Rearrange the way you want.  Then repeat process at step 1.

 

The 'initconfig' is necessary because if you don't do that, if you looked at the Main page, it would complain about missing disks (because the previous config was saved by unraid in step 2).

 

By unassigning the parity disk, we are just simplifying things & preventing possible loss of parity data if you do something wrong.  As long as parity is good, and you haven't written anything to the data drives, you will be able to recover data if you have to.

OK. That makes sence. I'll unasign the parity, reboot. run initconfig and get back with the results. Do I need to reboot to generate a new syslog after initconfig comand?

 

OK. Here's the latest syslog. After initconfig and Start, all drives show up green and none are unformatted. The Parity is still unassigned.  Should I reassign the parity drive now and run: mdcmd set invalidslot 99

 

Thank you again for the fast help!

 

Lev

syslog-2010-06-22_c.txt.zip

Well....

 

I went through all the steps prior to reassigning the parity and all the drives appeared green with no errors. Then I reassigned parity and did the following:

1. Install parity & assign via Devices page.

2. Start server & look on Main page, array should not Stopped.

3. From telnet or console type these commands:

initconfig

4. Refresh the Main page, array status should show "all new drives".

5. From telnet or console type this:

mdcmd set invalidslot 99

6. Click 'Start' button on Main - array should show up with all 'green' drive status.

7. Run a parity-check for the fun of it.

 

 

Now 4 unformatted drives  ???

 

Here is the syslog.

syslog-2010-06-23_a.txt.zip

From syslog, the lines:

Jun 23 06:37:26 Tower logger: mount: wrong fs type, bad option, bad superblock on /dev/md6,

Jun 23 06:37:26 Tower logger: mount: wrong fs type, bad option, bad superblock on /dev/md1,

Jun 23 06:37:26 Tower logger: mount: /dev/md2: can't read superblock

Jun 23 06:37:26 Tower logger: mount: /dev/md7: can't read superblock

 

show errors reading disk1, disk2, disk6, and disk7.

 

Try this: In the current state, unassign parity, initconfig, and then Start array - do you see same unformatted disks?

 

If yes, swap say disk1 with disk3, initconfig, then Start array - do you see same unformatted disks (if so, then problem is not physical drive that was in disk1, problem is with cable/backplane/controller port).

I unassigned the parity, ran initconfig and after Starting again the array all drives show up [glow=green,2,300]green[/glow]. None are unformatted. I am at work, so I cannot physically swap drives till later in the evening when I get home. However, if I do not see errors now, what does that mean?

 

Lev

 

(Used to be a "Restore" button that did this, but Joe made me take it off, but I'm going to put it back next release.)

I never asked it be removed.  I've asked for years that it be re-named.    We've just not been able to settle on a different name.

 

If you put a "button" back, at least re-name it to

"Initialize Configuration"

and change the affiliated description to indicate it "Sets a new initial configuration and immediately invalidates any prior parity calculations based on any previous configuration."  State the the new configuration will include only the currently assigned and working drives. 

 

Also, update your on-line manual in the wiki to reflect whatever you finally name it.  Otherwise, users will complain when the manual does not match what they are seeing on the screen.

 

Joe L.

 

I agree with Joe 

 

I don't remember the suggestion to take the button off. (maybe relocate it).

For years the board members have been asking for it to be renamed, (not removed).

I agree with Joe.  "Initialize Configuration" is better.

The text is a good suggestion too.

 

On my Thecus 5200BR, if you made a change to the array, you had to remove the array, then create the array.

On my ReadyNAS it was similar too.

 

Initialize Configuration is a good name for the function in unRAID.

Personally that button is asking for trouble. With my limited knowledge of the web interface I was able to get a few things setup for testing. I'm almost sure I would of hit that button and screwed things up even knowing better since I've read it here many times.

 

I did have one issue when I needed to find it, but did a search and found initconfig.

 

Personally keep it simple and if need be build a help command into the console and when you type "help" have it give you a list of useful commands that are out of the web config or are nice to know commands that somebody might actually forget.

 

Back in the day I used to use a lot of alias's and one of them was actually "Help" it would spit out a list of commands that often forgot. Maybe it was Help Me. I don't rember any more.

I unassigned the parity, ran initconfig and after Starting again the array all drives show up [glow=green,2,300]green[/glow]. None are unformatted. I am at work, so I cannot physically swap drives till later in the evening when I get home. However, if I do not see errors now, what does that mean?

 

Lev

 

 

I'd like to see this syslog.

 

It would appear that,

- having parity included in the array, causes other drives to get errors

- having parity not included, everything is fine

Agreed?

 

If so, implies something is wrong with parity disk, backplane, cable, or port.  Next step would be to physically install parity in a different slot, assign Parity to that slot, and see if there are errors.  If no errors then put a data disk in the old parity slot and see if you now get errors.

I unassigned the parity, ran initconfig and after Starting again the array all drives show up [glow=green,2,300]green[/glow]. None are unformatted. I am at work, so I cannot physically swap drives till later in the evening when I get home. However, if I do not see errors now, what does that mean?

 

Lev

 

 

I'd like to see this syslog.

 

It would appear that,

- having parity included in the array, causes other drives to get errors

- having parity not included, everything is fine

Agreed?

 

If so, implies something is wrong with parity disk, backplane, cable, or port.  Next step would be to physically install parity in a different slot, assign Parity to that slot, and see if there are errors.  If no errors then put a data disk in the old parity slot and see if you now get errors.

 

Easy enough. I agree, throwing parity into the mix screws things up. I will post later on tonight after I swap the drives.

 

Again, thank you for the continued support.

 

Lev

 

By the way, do you want me to run a syslog before I reboot to swap out the drives? reboot, run syslog, and then swap drives; or, forget about syslog and go straight to swapping?

 

 

I want to see syslog before swap - even though drives don't show 'unformatted', there could still be errors.

I want to see syslog before swap - even though drives don't show 'unformatted', there could still be errors.

 

OK. Do I need to reboot prior to generating syslog?

 

 

I want to see syslog before swap - even though drives don't show 'unformatted', there could still be errors.

 

Here is the syslog without rebooting.

 

If I add parity disk back in without rebooting, the other drives all appear green, and the message about invalid parity shows up. I'll swap out the parity drive to a different spot later on and post a new syslog.

syslog-2010-06-23_b.zip

In order to physically move hard drives around you must power down.  There are still errors in that system log.

I swapped the disk 1 and parity hard drives around, ran initconfig and mdcmd set invalidslot 99... Again disks 1, 6 and 7 are unformatted. I am getting very frustrated... 

syslog-2010-06-23_c.zip

I would not call this progress, but I've changed the Supermicro SAS cards out, and rebooted. After adding parity back in, running initconfig, and then mdcmd set invalidslot 99, all but disk 6 "look" OK, with disk 6 having errors. So, what's next?  ???

syslog-2010-06-23_d.zip