SimonAG Posted September 18, 2019 Share Posted September 18, 2019 (edited) I am about 16 months in to my unraid setup. I bought 4x6tb Seagate NAS drives 16 months ago when I bought my unraid hardware and I also had 4x4tb Seagate non-nas drives. All four of the 4tb drives are 4 years old with one approaching 5 years old. The 5 year old one has just come up with Reported Uncorrect Smart error at 50. Ideally I would replace all four but if i can get away with replacing one at a time that would be ideal cost wise. Parity is currently 90 days old. Should I run a parity check again before replacing the drive with a new 6tb drive? (parity is 6tb) or not? Also would you think it is best to replace all 4x4tb drives given their age? Edited September 18, 2019 by Simontv Quote Link to comment
Frank1940 Posted September 19, 2019 Share Posted September 19, 2019 (edited) I would say that that most folks are not worried about using non-NAS drives in Unraid. There is no evidence to support that they are any more prone to failure than the NAS drives that are being sold. (If you really look at things, failure rates of hard drives are actually fairly low. Certain models when looked at on a historical perspective are more likely to fail than other models but that data is usually too late to be of any real benefit to the purchaser.) You should always run a non-correcting parity check before you voluntarily replace a parity drive! If there are problems with this parity check, that is the time to address them! (90 days between parity checks is probably much too long an interval. Most folks are using monthly checks.) You should also have the Notifications setup to provide with warnings of problems. Settings >>> Notifications 11 hours ago, Simontv said: The 5 year old one has just come up with Reported Uncorrect Smart error at 50. Exactly what is the attribute number of this error? And is "50" the count? You could assist us by providing a Diagnostics report Tools >>> Diagnostics 11 hours ago, Simontv said: Also would you think it is best to replace all 4x4tb drives given their age? Not really! Google 'bathtub failure curve' Hard drives being a complex electronic-mechanical system have this type of failure mode. Most folks basically only replace a drive when it fails (or is starting to have issues) or when they run out of capacity. Age is no great indicator that a drive is going to fail. Any random drive could fail after 2 hours of use or after ten years of use. (By the way, your idea of purchasing drives over a period of time is good but not necessary for the reason you mentioned. Doing so will prevent you from buying into a lot of drives with poor initial quality due to a manufacturing problem. It also avoids shipping problems where the package is ends up being treated like a football. One drive in either situation is better than four.) You might want to have a spare drive on hand. Tested and ready to install. (Meaning 70 to 100 hours using Manufacturer's software or the preclear plugin/script) As a point of information, there is a procedure to replacing a 'failed' data disk with the parity disk while adding a larger parity disk at the same time: https://wiki.unraid.net/The_parity_swap_procedure#targetText=This procedure is strictly for,building parity will immediately begin. EDIT: I just realized that the Parity swap information is not really needed in your case as you have a 6TB parity drive. Personally, I would try to avoid this situation but it is not always possible. It takes a long time to complete and there is a possibility of something happening or going wrong... Edited September 19, 2019 by Frank1940 Quote Link to comment
SimonAG Posted September 19, 2019 Author Share Posted September 19, 2019 (edited) I know I should have done more but I have some heat issues in my current case and it is worse when I run a parity check. I have been waiting until the temperature cools down a bit to run it again. I have been meaning to address the heat issue as well by buying a small rack and fixing ventilation. Problem is I think the case design is prone to high temps no matter what I do. Which was clear early on. I don't think the heat caused the smart errors though as disk 7 was always at normal temp, it is just disk 2 and 3 that get hot. How crucial is it to run a parity check before swapping out a drive.? I do need some extra space, down to 8tb now. I bought two 6tb drives to replace two of the drives approaching 5 years and the other two that are 3 years 9 months I will leave until I run out of space again or they have issues. I see in the syslog disk 7, read errors. I will replace that disk right away. Is it worth parity checking or would I be better off just replacing it and see if it rebuilds ok, I assume if it doesn't then I could just plug back in the "faulty" drive and try a parity check or copy off the data? i have attached the logs. Edited September 26, 2019 by Simontv Quote Link to comment
Frank1940 Posted September 20, 2019 Share Posted September 20, 2019 Looking at the log, I would replace disk 7 at this point without trying a parity check. (Those read errors would probably cause the parity check to fail in any case.) Hang onto disk 7 in case, you do have any problems rebuilding it. Quote Link to comment
SimonAG Posted September 20, 2019 Author Share Posted September 20, 2019 Thank you for your help Frank! I will report how i get on. Quote Link to comment
SimonAG Posted September 26, 2019 Author Share Posted September 26, 2019 I replaced the drive and rebuilt from parity without issue, thanks for your help. Now just have to solve the temperature issue so i can schedule the parity check! 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.