Unraid doesn't like one disk...

[thegizzard]

I can boot unraid without this disk plugged in.  If I plug in this disk, the system hangs and doesn't complete a boot.  I can't even get a syslog out of it.  When I insert the disk after a clean boot without it, it kicks off a bunch of kernel errors and stops unraid.  I can't even get a syslog out of it. 

This disk used to work just fine in my array, no issues... although it is one of those dreaded 3 TB Seagates that have failed so many times for me. 

 

Can anyone think of why just plugging in a disk (into any slot btw) kills unraid?

 

 

[RobJ]

I'd be worried about shorted leads, or other physical damage.  I would certainly NOT attach it to my computer!  Most damage stops a device from working correctly, or not at all.  Electrical damage likes to be shared, to pass it on, damaging whatever is connected!

[trurl]

I've not had this with unRAID and any disk I have tried so far, but I have had a faulty drive prevent Windows from booting. Stay away from that disk.

[thegizzard]

Darnit.  ok.  so follow up question.

 

11 disk array.  one redball (3tb).  new disk (5tb) pre-cleared and set to go.  second disk (3b) fails before I rebuild array with new disk.  (second disk is the aforementioned 3TB Seagate)

 

What's the best approach to rebuild the array with two missing disks?

 

Actually now I am thinking the original problem could be the port the redball was plugged into.

[trurl]

Darnit.  ok.  so follow up question.

 

11 disk array.  one redball (3tb).  new disk (5tb) pre-cleared and set to go.  second disk (3b) fails before I rebuild array with new disk.  (second disk is the aforementioned 3TB Seagate)

 

What's the best approach to rebuild the array with two missing disks?

 

Actually now I am thinking the original problem could be the port the redball was plugged into.

If you only have one parity drive then there is no rebuild with two missing. There may be more elaborate methods to recover some or all of your files though. Is the 2nd disk actually redball? I don't think unRAID will try to write to any failed disk if there is more than one missing since it is no longer possible to emulate. Post a diagnostic, then check your connections.

[trurl]

Sorry, I get it, the 2nd disk is the one you can't boot with. Post a diagnostic anyway.

[trurl]

Might be possible to fool unRAID into rebuilding the 2nd disk instead. How long was that other one redballed?

[trurl]

And don't do anything without further instructions unless you have good backups or don't care about the contents of those 2 disks.

[thegizzard]

Could I try and clone the 2nd disk and then rebuild from the clone?

 

Sent from my SM-N920V using Tapatalk

 

 

[trurl]

Could I try and clone the 2nd disk and then rebuild from the clone?

 

Sent from my SM-N920V using Tapatalk

I thought the "2nd disk" was the one you couldn't boot with.

[trurl]

Would still like a diagnostics if you can get it.

[trurl]

... How long was that other one redballed?

And an answer to this question.

[thegizzard]

I can't. But maybe I can clone it with another PC then try the cloned drive

 

Sent from my SM-N920V using Tapatalk

 

 

[trurl]

I can't. But maybe I can clone it with another PC then try the cloned drive

 

Sent from my SM-N920V using Tapatalk

Do you know how to clone a disk? If that would work it would be preferable to what I was going to propose, but it seemed as if the disk is unusable. Go ahead and try to clone it on another computer.

 

And while you're trying that, can you provide me with what I've asked? I have asked for diagnostics and also asked how long the other drive was redballed. Those can help decide what to do in case that drive actually turns out to be unusable.

 

[thegizzard]

Ok.  I have attached diagnostics.  Redball was for several weeks.  Maybe a month.

 

I have the following which I think I can use to clone the drive. 

 

https://amzn.com/B00IKC14OG

 

But I have not tried to use it for that yet, and definitely not for a linux file system.  I need to buy a second replacement and then look into the clone.

 

All of this said, I am waiting on replacement hardware which will be here next week.  My plan was to put all these drives on the new system next weekend, but I wanted to sort out the redball first.

 

 

 

megatron-diagnostics-20160908-2125.zip

[trurl]

Just compiling some facts for future reference.

 

From your syslog

unRAID System Management Utility version 6.2.0-rc4

disk0: (sdm) WDC_WD5001FZWX-00ZHUA0_WD-WX21D955YN7A
disk1: (sdg) Hitachi_HDS5C3030ALA630_MJ1311YNG2A3UA
disk2: (sdj) TOSHIBA_DT01ACA300_Z5HPRLDGS
slot: 3 empty
disk4: (sdi) Hitachi_HDS722020ALA330_JK1130YAGKSS4T
slot: 5 missing
disk6: (sdh) ST4000DM000-1F2168_S300JSDS
disk7: (sdc) Hitachi_HDS722020ALA330_JK1130YAGKTUKT
disk8: (sdn) HGST_HDN724040ALE640_PK2334PAKTRVBT
disk9: (sdf) ST3000DM001-1CH166_W1F22FBJ
disk10: (sdd) WDC_WD20EARS-00MVWB0_WD-WMAZ20039357
disk11: (sde) ST4000DM000-1F2168_S300JRKG
slot: 29 empty
cache device: TOSHIBA_DT01ACA100_26JHP60FS (sdk)
flash device: SanDisk_Cruzer_Edge_200608760202A3C1C219-0:0 (sda)

 

What disk is this?

TOSHIBA_HDWE150_76IAK972F57D (sdl)

 

SMART OK except for this which we can't do anything about for now

Serial Number:    WD-WMAZ20039357
197 Current_Pending_Sector  0x0032   200   200   000    Old_age   Always       -       1

 

Which slot was redball and which was the "unbootable"?

[trurl]

Redball was for several weeks.  Maybe a month.

Definitely not what I wanted to hear. That means the disk may be very out-of-sync and I was hoping we could use it to rebuild the other one.

 

Do you think it was written to? unRAID will not have actually written to the physical disk, but it will have written to the emulated disk if anything tried to write to that disk and that would have made parity and the physical disk out-of-sync.

[thegizzard]

 

What disk is this?

TOSHIBA_HDWE150_76IAK972F57D (sdl)

 

HDWE150_76IAK972F57D is the 5TB disk I bought to replace the original redball.

 

Which slot was redball and which was the "unbootable"?

 

disk 3 - Z1F2PLM3 was the original redball... see the attached capture.

disk 5 - W1F0ZQ9M is the disk i cannot boot with now.

 

The one thing I did not do was confirm which disks were in each slot in the chassis before I started mucking around.  So now I am unsure if there may have been a problem with the disk 3 slot, that has affected disk 5 now.  Either way, I have a new 3TB disk on the way that I will use to attempt to clone W1F0ZQ9M.

 

Unless you tell me otherwise, I will wait until the new MB and Dell PERC H310 card arrives before I try to reboot the array.

 

Do you know if there is a Windows utility that will allow me to read the linux file system if I attach it via my USB enclosure?  I am hoping W1F0ZQ9M  is readable outside of UNRAID.

 

Also, do you know if there is something I can use to test each port on the enclosure before I put the whole system back together?

[trurl]

Unless you tell me otherwise, I will wait until the new MB and Dell PERC H310 card arrives before I try to reboot the array.

 

Do you know if there is a Windows utility that will allow me to read the linux file system if I attach it via my USB enclosure?  I am hoping W1F0ZQ9M  is readable outside of UNRAID.

 

Also, do you know if there is something I can use to test each port on the enclosure before I put the whole system back together?

The main rule of troubleshooting is only change one thing at a time. In other words, changing your other hardware in the middle of this problem is not a good idea unless we determine that your other hardware is at fault. We can discuss that later if it does come to that.

 

Don't have any personal experience with reading ReiserFS on Windows. You might try a google search for "read reiserfs from windows" or something like that.

 

Don't know of any specific SATA port tester. Have you tried looking for each drive on each port in your BIOS?

[thegizzard]

I agree. This was not my plan at all. It started with my MV8 cards going bad and then realizing I need to upgrade from PCI-X. 

 

Ok.  So for now my plan is to load unRAID onto an old core 2 duo PC and attached the bad drive via USB.  If it can be read I will clone it and try and drop the clone in Megatron. 

 

My most critical files are on CrashPlan so I know I can get the necessary stuff back if it's a total catastrophe.  I think I can also copy each disk one at a time to a clean array if I needed to, right? 

 

 

Sent from my SM-N920V using Tapatalk

 

 

[trurl]

My most critical files are on CrashPlan so I know I can get the necessary stuff back if it's a total catastrophe.  I think I can also copy each disk one at a time to a clean array if I needed to, right? 

Glad to hear you have backup for criticals.

 

You could copy each disk one at a time to a clean array, or you could just put the disks themselves in a clean array and start from there.

[thegizzard]

If I put the disks in a clean array and start from there, then the only data lost will be on the disks missing from the array correct?  In other words if disks 3 and 5 are missing and restart the array the only data lost will be from disks 3 and 5?

 

Sent from my SM-N920V using Tapatalk

 

 

[trurl]

If I put the disks in a clean array and start from there, then the only data lost will be on the disks missing from the array correct?  In other words if disks 3 and 5 are missing and restart the array the only data lost will be from disks 3 and 5?

 

Sent from my SM-N920V using Tapatalk

That's correct. With unRAID each disk is an independent filesystem. This is probably the best feature of unRAID that we all take for granted. Not only does it mean that you won't lose all your data, it also means you can mix drive capacities.

[thegizzard]

Great.  Then I am not going to freak out too much yet.  I think I can live with the worst case scenario if it come to that.

[trurl]

Just one more comment about putting disks with data on them into a new array. They must all be put in before parity is built or unRAID will want to clear them to keep parity valid.