swells Posted February 25, 2016 Share Posted February 25, 2016 Ok, I need some help. I picked up a new 8TB Seagate drive to serve two purposes. I have a few 2TB WD20EARS drives still in play and one of them has 1 sector pending relocation. I also am getting low on space and need to expand soon. The 8TB drive will become my new parity drive to replace a 3TB drive in use currently and the 3TB parity drive will replace this 2TB WD drive so I can retire it (it has been in use for ~5 years in this server). So that is the background. The drive was finishing up a single pass of preclear on the post read phase and was about 70% complete last night. I was about to go to bed and figured I would kick off a parity check and make sure all was well before the preclear finished and I could continue on. I woke up to the preclear complete, but the 8TB disk was spun down and would not spin up. It was basically not available any longer. I checked the logs and it was full of errors. Looks like two drives had some issues, ATA12 is the 2TB WDEARS drive I would like to replace, it reconnected to the controller and is fine now from what I can tell. And ATA15 (sdn) is the 8TB Seagate and it did not and is offline. I have no idea what I should be looking at. Is there a huge problem here? Did i over stress something by running a parity check while running the preclear? If someone could please take a look at the logs you will see everything of relevance I believe. I have removed a bunch of crap that was spamming the logs, which I have also never seen before, but it was basically thousands of lines of this: Feb 24 23:47:49 Tower kernel: program smartctl is using a deprecated SCSI ioctl, please convert it to SG_IO Errors.txt Quote Link to comment
trurl Posted February 25, 2016 Share Posted February 25, 2016 Instead of an uncompressed and incomplete syslog, for V6 you should always go to Tools - Diagnostics and post the complete diagnostics zip file. My guess is that you have an unreliable SATA or power connection on one or more drives caused when you were poking around inside. Stop, shutdown, check all connections, reboot, then post diagnostics. Quote Link to comment
swells Posted February 25, 2016 Author Share Posted February 25, 2016 Ok, here you go. tower-diagnostics-20160225-0847.zip Quote Link to comment
trurl Posted February 25, 2016 Share Posted February 25, 2016 Ok, here you go. Now give us another one after you have checked your connections and rebooted. Quote Link to comment
JorgeB Posted February 25, 2016 Share Posted February 25, 2016 Before running a parity check I would run an extended SMART test on the WD20EARS with the pending sector, if it doesn’t pass you have to replace it before any other upgrades. Quote Link to comment
swells Posted February 25, 2016 Author Share Posted February 25, 2016 I just ran an extended. It still passes. Has a current pending sector count of 1 and offline uncorrectable count of 1. Otherwise no errors. Quote Link to comment
JorgeB Posted February 25, 2016 Share Posted February 25, 2016 Post a new diagnostics after rebooting, Seagate had dropped offline and there’s no SMART report on previous one. Quote Link to comment
swells Posted February 25, 2016 Author Share Posted February 25, 2016 Here is a new one after a reboot and reseating cables. I didnt have time before work to do a complete rip apart and check but I can if I need to later tonight. Nothing sticks out to me in the logs. tower-diagnostics-20160225-1036.zip Quote Link to comment
JorgeB Posted February 25, 2016 Share Posted February 25, 2016 I just ran an extended. It still passes. Has a current pending sector count of 1 and offline uncorrectable count of 1. Otherwise no errors. Device Model: WDC WD20EARS-00MVWB0 Serial Number: WD-WMAZA4049015 Num Test_Description Status Remaining LifeTime(hours) LBA_of_first_error # 1 Extended offline Completed: read failure 90% 39715 291445898 SMART test failed, you have to replace this disk first. 8TB Seagate looks fine, maybe bad connection. Quote Link to comment
swells Posted February 25, 2016 Author Share Posted February 25, 2016 Ok, thanks! I guess I was just looking at the part that said passed up top. Any way for me to do this according to my original plans without purchasing another drive? Quote Link to comment
JorgeB Posted February 25, 2016 Share Posted February 25, 2016 Ok, thanks! I guess I was just looking at the part that said passed up top. Any way for me to do this according to my original plans without purchasing another drive? https://lime-technology.com/wiki/index.php/The_parity_swap_procedure But before trying that I would run another preclear on the Seagate or an extended SMART test to make sure it’s good. Quote Link to comment
JonathanM Posted February 25, 2016 Share Posted February 25, 2016 Ok, thanks! I guess I was just looking at the part that said passed up top. Any way for me to do this according to my original plans without purchasing another drive? Make sure any files you don't want to lose are backed up before you mess with it more. There is a not insignificant risk of data loss if something doesn't go just right. Quote Link to comment
swells Posted February 25, 2016 Author Share Posted February 25, 2016 Nothing I can't afford to lose on this server. Entirely media. I would however rather be safe than sorry, so will probably try to replace that disk before I do anything else. I do however wish to continue with my move to bigger 8TB drives. Any problems with the below plan? With the last couple steps, is there a way to expand the file system once the larger parity drive is in place or do I need to remove it and put it back in like a new disk to utilize the space? My current plan is something like: 1) Ensure 8TB disk is fine 2) Rebuild bad 2TB disk onto precleared 8TB disk. 3) Order another 8TB disk and preclear it. 4) Replace current 3TB disk with new 8TB disk. 5) Replace 8TB data disk previously used to replace failed 2TB disk with 3TB former parity disk. 6) Utilize full 8TB of original 8TB disk in array. Quote Link to comment
JorgeB Posted February 25, 2016 Share Posted February 25, 2016 My current plan is something like: 1) Ensure 8TB disk is fine 2) Rebuild bad 2TB disk onto precleared 8TB disk. 3) Order another 8TB disk and preclear it. 4) Replace current 3TB disk with new 8TB disk. 5) Replace 8TB data disk previously used to replace failed 2TB disk with 3TB former parity disk. 6) Utilize full 8TB of original 8TB disk in array. You can't do this, parity has to be same or larger size than any data disk, you can do the parity swap procedure I linked above, it will first copy your parity data to the 8tb disk, then rebuild the bad 2tb on the old parity. Quote Link to comment
swells Posted February 25, 2016 Author Share Posted February 25, 2016 You can't do that and just not utilize the extra disk space? hmm, well crap. Quote Link to comment
JonathanM Posted February 25, 2016 Share Posted February 25, 2016 Nothing I can't afford to lose on this server. Entirely media. I would however rather be safe than sorry, so will probably try to replace that disk before I do anything else. I do however wish to continue with my move to bigger 8TB drives. Any problems with the below plan? With the last couple steps, is there a way to expand the file system once the larger parity drive is in place or do I need to remove it and put it back in like a new disk to utilize the space? My current plan is something like: 1) Ensure 8TB disk is fine 2) Rebuild bad 2TB disk onto precleared 8TB disk. 3) Order another 8TB disk and preclear it. 4) Replace current 3TB disk with new 8TB disk. 5) Replace 8TB data disk previously used to replace failed 2TB disk with 3TB former parity disk. 6) Utilize full 8TB of original 8TB disk in array. Unraid will not allow you to use a disk that appears larger than the current parity drive as a data drive. In order for this plan to work, you would need to use a HPA to block off all but 3TB of the 8TB drive so it will appear the same size as your current parity. This procedure is riskier to implement than the parity swap that johnnie.black linked IMHO. Quote Link to comment
swells Posted February 25, 2016 Author Share Posted February 25, 2016 Alright, i'll run with that then. Should I do a full preclear again or just a post read? Thats where it was when it disconnected from the controller last night. As far as I can tell, everything was ok with the preclear up to that point. With the plug in it looks like I can just do the post read and save a couple days, but I have no problem doing it all again if need be. Just not sure about leaving that bad disk there a while longer. Quote Link to comment
JorgeB Posted February 25, 2016 Share Posted February 25, 2016 You can't do that and just not utilize the extra disk space? hmm, well crap. No, it’s not possible, and I don’t see I you would want to do that, you can just: -do the parity swap -buy another 8tb disk and add it to the array Unless I’m missing something you’ll be with the same setup as above with less steps Quote Link to comment
JorgeB Posted February 25, 2016 Share Posted February 25, 2016 Alright, i'll run with that then. Should I do a full preclear again or just a post read? Thats where it was when it disconnected from the controller last night. As far as I can tell, everything was ok with the preclear up to that point. With the plug in it looks like I can just do the post read and save a couple days, but I have no problem doing it all again if need be. Just not sure about leaving that bad disk there a while longer. Full precleasr would be better, but it will also leave you more time with a bad disk, that leaves you vulnerable in case of another disk failure, so in my view it’s a toss up. Quote Link to comment
swells Posted February 25, 2016 Author Share Posted February 25, 2016 Yeah, honestly this drive may have been bad for a while. Who knows. It is one of the oldest in there and who knows how often it is accessed. I will preclear again and go from there. Thanks for your help! Quote Link to comment
tdallen Posted February 25, 2016 Share Posted February 25, 2016 How full are your drives? Do you have the option of moving data around to come up with a spare disk? Quote Link to comment
swells Posted February 25, 2016 Author Share Posted February 25, 2016 I have about 700GB left I think. So maybe 100-200GB tops on any given disk. Quote Link to comment
swells Posted February 27, 2016 Author Share Posted February 27, 2016 Ok. So I ran the post read portion of the preclear and it was fine. So I trusted the new 8TB disk and wanted to replace that failed 2TB disk asap. I did the swap and everything went fine. After the swap I started a nocorrect parity check and so far, at 27% complete (2.2TB or so), there are 4201 sync errors. My first thought is to let this run and then do it again and see if it is the same. If it is, do I correct it? How should I proceed at this point? Quote Link to comment
JorgeB Posted February 27, 2016 Share Posted February 27, 2016 Assuming all your disks are healthy you should run a correcting parity check, you probably have some sync errors caused by the bad disk. Quote Link to comment
swells Posted February 28, 2016 Author Share Posted February 28, 2016 Should I skip the 2nd nocorrect check and maybe run extended smart tests on my other data drives to ensure they are healthy? Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.