March 30, 20179 yr I logged into my unRaid webGui tonight to find that one of my drives has been disabled. I've run the diagnostic tool as suggested in other threads, but I'm not entirely sure how to interpret the SMART test results. Would anyone be willing to take a look at the output of my diagnostic test and interpret what's there? I'd be grateful. The disk that is currently emulated is my Disk 8. It's serial is HGST_HDS5C4040ALE630_PL1331LAGESD6H. Thanks, in advance, for any input you can provide . media-diagnostics-20170329-2126.zip
March 30, 20179 yr Looks like you rebooted after you saw that it was disabled, which unfortunately means that the logging that preceded it being disabled (due to a failed write) is gone. SMART on the drive is clear, so presumably the drive dropped offline for some reason (power glitch / cabling issue). Right now you've pretty much got no options but to rebuild onto itself (stop array, unassign the disk, start array, stop array, reassign the disk, start array -> a rebuild will commence)
March 30, 20179 yr Author Squid ... thanks for the quick reply. I'm not really sure why I rebooted. I guess I was hoping a good old-fashioned reboot might make things better. I'll start a rebuild now and cross my fingers
March 30, 20179 yr Author If the rebuild is successful am I in the clear? Or, is there something that I should look for during/after the rebuild that will indicate the disk needs replacement?
March 30, 20179 yr Community Expert 1 minute ago, jmos1277 said: I guess I was hoping a good old-fashioned reboot might make things better. Sometimes a reboot might reset something in the hardware, but it won't make anything different in the software, and the software had already decided your disk's contents were invalid and needed to be rebuilt.
March 30, 20179 yr Just now, jmos1277 said: I guess I was hoping a good old-fashioned reboot might make things better. Once disabled, always disabled until rebuilt... Reboots on any OS all do the exact same thing. They get rid of the error / bug / weird funky thing happening until it happens again.
March 30, 20179 yr Community Expert Just now, jmos1277 said: If the rebuild is successful am I in the clear? Or, is there something that I should look for during/after the rebuild that will indicate the disk needs replacement? SMART looked OK. unRAID monitors certain critical SMART attributes and will put a warning in the webUI if they begin to show trouble. And even better, Notifications will tell you about these and other sorts of problems without you even looking at the webUI, or even being in the same place as your server. You absolutely must setup Notifications if you haven't already.
March 30, 20179 yr 1 minute ago, jmos1277 said: If the rebuild is successful am I in the clear? Or, is there something that I should look for during/after the rebuild that will indicate the disk needs replacement? What you're going to be looking for ( or what if you hadn't rebooted what we would have been looking for ) is errors logged during the rebuild: Things like ATA Reset errors, etc. Stuff like that would tend to imply a slightly loose cable / power problem. Or it also is possible that the drive just burped and had a brain-fart and decided to go on vacation for a couple of minutes. Either way, if errors are logged during the rebuild, or if a drive gets disabled in the future, diagnostics before rebooting will usually tell us the probable reason why...
March 30, 20179 yr Author trurl, thanks for the suggestion. I thought I had notifications setup. However, upon further inspection I discovered that I only had "Array Status" notifications emailed to me. All other notifications I only had setup as browser notifications. That has been remedied.
March 30, 20179 yr Author 2 minutes ago, Squid said: What you're going to be looking for ( or what if you hadn't rebooted what we would have been looking for ) is errors logged during the rebuild: Things like ATA Reset errors, etc. Stuff like that would tend to imply a slightly loose cable / power problem. Or it also is possible that the drive just burped and had a brain-fart and decided to go on vacation for a couple of minutes. Either way, if errors are logged during the rebuild, or if a drive gets disabled in the future, diagnostics before rebooting will usually tell us the probable reason why... Thanks. I'll keep an eye on the rebuild and its results. In the meantime, I'm buy another disk just in case things go south. And, if my disk turns out to be okay, then I'll just use the new disk as my second parity drive. I've been meaning to set that up for a while now.
March 30, 20179 yr Author UPDATE: My rebuild finished just moments ago. I'm not home so I cannot access my logs in detail. However, my notification email indicated that my Disk 8 has returned to normal operation and that my rebuild completed with 0 errors. It seems everything is looking okay for now. Event: unRAID Parity sync / Data rebuild Subject: Notice [MEDIA] - Parity sync / Data rebuild finished (0 errors) Description: Duration: 14 hours, 50 minutes, 32 seconds. Average speed: 74.9 MB/s Importance: normal
March 30, 20179 yr Author Made it home and checked my logs. I don't see anything suspicious, so I think I'm in the clear for now. I'm currently running an extended SMART test on the drive. Seems like the test is going to take forever. I guess the WD-RED that I ordered from newegg is going to become a second parity drive. A fortunate result :-) Thanks again to all for taking the time to comment and help me out.
June 5, 20179 yr Author I just received an email from my unRAID box that it has encountered a disk error again. Once again, it appears that my disk8 has been disabled. When this error occurred 3 months ago, I rebuilt the data onto the same physical disk without replacing it. Everything seemed to be okay for the past few months but now the disk has been disabled again. I have not reboot my system (a mistake I made last time) and was able to grab my syslog and a diagnostics dump. Are there any unRAID sleuths willing to take a look and provide some expert insight? I see the following types of errors (most of them are repeated many times): kernel: sas: sas_eh_handle_sas_errors: task 0xffff8801a197ba00 is aborted kernel: sas: sas_ata_task_done: SAS error 8a kernel: ata22.00: failed to IDENTIFY (I/O error, err_mask=0x11) kernel: blk_update_request: I/O error, dev sdu, sector 1792104 kernel: md: disk8 read error, sector=1792040 emhttp: err: mdcmd: write: Input/output error kernel: md: do_drive_cmd: disk8: ATA_OP e0 ioctl error: -5 syslog.txt media-diagnostics-20170605-0259.zip
June 5, 20179 yr Community Expert Same disk dropping twice points to a disk problem, but since it's on a SASLP and these are know to sometimes drop disks you can try and confirm, because it dropped offline there's no SMART report, but assuming that it looks good you can swap that disk with one on the onboard Intel controller, if it drops again there replace it.
June 5, 20179 yr Author 1 minute ago, johnnie.black said: Same disk dropping twice points to a disk problem, but since it's on a SASLP and these are know to sometimes drop disks you can try and confirm, because it dropped offline there's no SMART report, but assuming that it looks good you can swap that disk with one on the onboard Intel controller, if it drops again there replace it. johnnie.black, thanks for the reply. To clarify, I assume you mean that I should physically swap the disk with one currently connected directly to my motherboard, is that correct? That is, shut down server, physically move disks, and restart? Can/Should I swap the disks prior to rebuilding the currently disabled disk? If the disk drops again when it's connected to the Intel controller, I guess that points to a bad disk. If the disk that I swapped with ends up disabled in a few months time, I guess that points to a bad SASLP. Does is all sound about right?
June 5, 20179 yr Author 1 minute ago, johnnie.black said: Yes on all questions Excellent. Thanks again for the reply. I'll start the process in the morning after my current parity check completes. Any recommendations on replacements cards for my SASLP cards if they need replacement? Are the AOC-SAS2LP-MV8 cards any better?
June 5, 20179 yr Community Expert Avoid Marvell controllers with current unRAID, like both the SASLP and the SAS2LP, LSI is the best option, e.g., 9211-8i or clones.
Archived
This topic is now archived and is closed to further replies.