July 18, 201213 yr Hello, I am hoping someone can help me. I have a small UnRaid server, currently running the latest rc (8168 test), which has been running fine for the past 8 months or so. Recently, I had my drive 5 fail with the "disk_disabled" message, and a red ball. Thinking the drive was probably bad, I went and replaced the 2TB drive with a new 3TB WD green drive. I pre-cleared it (successfully), and then started rebuilding data. I shortly had the rebuild fail, with 1 write error. So I replaced the data cable, and chose a different power connection. Same issue - the rebuild starts, then fails with a single write error. So, thinking the 2-port sata card I have it connected to may be acting flaky, I added in a port replicator card (I'd thought I was replacing the actual card, but it turns out that the SyBase card I'd gotten simply connects to a single data port and gives you 2 ports from the one). Same issue, with a write error shorty after starting a data rebuild. So - I have replaced the drive, replaced the sata cable and the power connector, and am no longer using the sata port that I had been. I'm not sure where to go from here! If anyone has any ideas, I would really appreciate it. Thank you, Sean
July 18, 201213 yr Going to need a syslog. First thing that comes to mind is that your using a hotswap bay that could be faulty. That or you got unlucky and replaced something faulty with another faulty part.
July 18, 201213 yr Author Ok, thank you - I will create and post one when I get home from work. I CAN tell you that I do not have the drive in any kind of bay, unfortunately.
July 18, 201213 yr Author Hello, I am attaching a system log. The log shown in the browser was considerably longer than the one in the window that pops up under the "log" button, so I am attaching the entire thing. Apologies if it contains lots of extraneous information. ~Sean syslog.txt
July 20, 201213 yr Author Hello, I have now replaced the sata cabe again with a brand-new one. I also swapped the two ports on the card, to eliminate that as a variable. So now, I am having the same issue - start the data rebuild, and soon get a write error. The abbreviated system log: Jul 19 21:17:55 Tower kernel: md: recovery thread rebuilding disk5 ... Jul 19 21:17:55 Tower kernel: md: using 1536k window, over a total of 2930266532 blocks. Jul 19 21:24:53 Tower kernel: attempt to access beyond end of device Jul 19 21:24:53 Tower kernel: sdf1: rw=1, want=4734320648, limit=4294967295 Jul 19 21:24:53 Tower kernel: md: disk5 write error Jul 19 21:24:53 Tower kernel: handle_stripe write error: 4734320640/4, count: 1 Jul 19 21:24:54 Tower kernel: md: md_do_sync: got signal, exit... Jul 19 21:24:54 Tower kernel: md: recovery thread sync completion status: -4 Jul 19 21:24:54 Tower kernel: md: recovery thread woken up ... Jul 19 21:24:54 Tower kernel: md: recovery thread has nothing to resync Thank you, Sean
July 20, 201213 yr Post a smart report of disk5. I saw your posts earlier..didn't notice anything obvious except the write error...went to work, proceeding to forget about it. Edit: Oh yeah...there was this: Jul 17 20:56:27 Tower kernel: attempt to access beyond end of device Jul 17 20:56:27 Tower kernel: sdf1: rw=1, want=4734320648, limit=4294967295 File system check possibly? You rebuilt the new drive with this previous one... which likely had a fs error? Rebuilding wouldn't fix the problem, would copy the entire disk <including errors> and the fs would still need to be repaired. Edit2: http://lime-technology.com/forum/index.php?topic=17358.msg156916#msg156916
July 20, 201213 yr Author Hello, I apologize for the delay in my response - busy day at work! Anyway, I have attached the smart report fro disk 5. Also, I DO have the 2 TB drive here, just sitting on my desk. I am thinking of running a file system check, but will wait for your reply before doing so. It seems as though not much was written to my disk 5 before the write error came up (only takes like 5-10 minutes for it to happen), so I am not sure if it would be disk 5 that I would want to run it on? Thank you again for taking the time to look at this! ~Sean smart_status_report.txt
July 20, 201213 yr The Smart report looks fine (RAW values = 0) 197 Current_Pending_Sector 0x0032 200 200 000 Old_age Always - 0 198 Offline_Uncorrectable 0x0030 100 253 000 Old_age Offline - 0 199 UDMA_CRC_Error_Count 0x0032 200 200 000 Old_age Always - 0 The problem is now that you have a parity reconstruction of the old 2TB drive with fs errors. So if it tries to rebuild onto the 3TB, it is going to copy the fs error on the 3TB..hence why you get the "attempt to access beyond end of device" errors. Do you have a smart report from the 2TB drive before you removed it? Does the 2TB still have the files on it? If the 2TB does not indicate any smart errors... What I think should happen is this: 1. Re-install the 2TB drive in place of the 3TB 2. Run the filesystem check http://lime-technology.com/wiki/index.php?title=Check_Disk_Filesystems 3. Post either here or new thread with the results 4. Run commands per forum instructions... That would likely get 2TB operational again, assuming that there are no SMART issues (pending sectors etc) I believe that if you ran a filesystem fix on the 3TB, once you tried to rebuild it from parity it would create fs issues again. Although, there should be no issue or danger if you ran reiserfsck --check on the 3TB. <too many "thinks", "should" and "believes" for me to be 100% on these instructions. Hopefully someone else will affirm or correct the above.>
July 20, 201213 yr Author I do not have a smart report from the 2TB, unfortunately. I have installed it in place of the 3TB - but now, I am getting the message that the "Replacment disk is too small.", so the array won't start. Message next to the start button is "The replacement disk must be as big or bigger than the original."
July 20, 201213 yr SO...you have to keep the 3TB. You have a couple options <again, I've never run into this so...careful!> Do you have another port to plug the 2TB into <but not add it to the array> You could get a smart report that way....and possibly run reiserfs on that drive.
July 20, 201213 yr I edited my reply. Plug in the 2TB, not add it to the array <don't want to mess up the current parity>..and get the smart report.
July 20, 201213 yr If I haven't been clear...be very careful. I have never run into this specific problem, and don't want you to lose data. If the 2TB drive is not recoverable, I'm not certain that you'll be able to rebuild onto a new drive from the current parity...but I don't want to lose the current parity. It is the only hope if the 2TB is toast.
July 20, 201213 yr Author Oh, I see. I wasn't sure I could run a FS check o the drive if it wasn't part of the array. At the moment, I have disconnected the 3TB and plugged in the 2TB. I believe that the array thinks that it (and the 3TB) are new disks. But, I will try running a short smart test with unmenu on the 2tb. I don't have an extra port at the moment, but I could re-connect this port multiplier card I got the other day... all that card really seems to do is give you 2 ports out of 1 - it doesn't actually connect to the mb, you just plug one it into an existing sata port, and then you have 2 ports available on the card.
July 20, 201213 yr Believe you can run reiserfs --check...it will have to be unmounted anyways per the link I provided.
July 21, 201213 yr Author Ok. I realize that it needs to be unmounted, but as long as it can be "outside the array", as it were, I can't imaging a check would actually hurt it. Here's the status report, attached - I didn't let it run for more than 15 min or so. smart_report_2tb.txt
July 21, 201213 yr 197 Current_Pending_Sector 0x0032 200 200 000 Old_age Always - 0 198 Offline_Uncorrectable 0x0030 200 200 000 Old_age Offline - 0 All good on the 2TB as far as SMART goes.
July 21, 201213 yr Author Cool. That may be good news, then! Now - I am trying to check the fs, but I keep getting "Failed to open the device '/dev/md5': No such file or directory". The wiki link you gave me wants me to start the array in maintenance mode - but I can't start the array b/c the disk is too small.
July 21, 201213 yr You will have to put the 3TB drive back in, then add the 2TB. But not assign the 2TB Sent from my SAMSUNG-SGH-I897 using Tapatalk 2
July 21, 201213 yr Author Ok. Done. I am still getting the device not found error, though - I guess I'll have to upgrade the license to Pro?
July 21, 201213 yr Author Well - I have upgraded. Now I just have to wait for an email with the new key.
July 21, 201213 yr Author Ok. Got the key, rebooted, etc. Now - I can't seem to run a fs check on a disk outside of the array! Not sure what to do here.
July 21, 201213 yr What you could do is remove the 3TB drive. Assign the 2TB to the slot and run the new config utility. This will set a new disk configuration, and invalidate parity. I don't believe the current parity is good anyways. Then, you should be able to run the file system check. Sent from my SAMSUNG-SGH-I897 using Tapatalk 2
Archived
This topic is now archived and is closed to further replies.