Carrosive Posted August 31, 2017 Share Posted August 31, 2017 I've been running unRAID for a few days now and one of my drives is acting strange so I thought I would ask for some advice. I built a new server to run unRAID using three new 3TB drives and once up and running I copied all my files from my previous server (Ubuntu) over. After the files were all moved I then moved the four 3TB drives I had in my previous server into the new server. On moving the drives I installed a Supermicro AOC-SASLP-MV8 for the additional SATA connections, I also reconnected the two new data drives from the mobo (as they were when first setup in unRAID) to this card leaving the parity drive connected to the mobo. The new drives and the drives moved from my previous server were all picked up by unRAID and on starting the array the files copied over to the new drives were still accessible so I don't think there's an issue with the card. (?) All four drives moved from my previous server to my new server were all pre-cleared, formatted and successfully added to the array. SMART reports from short self-tests were all good and the four moved drives have been unused for the couple days that they have been in the new server. Today though I noticed that the last disk in the array was marked as faulty with the red X, and the SMART report was empty: smartctl 6.5 2016-05-07 r4318 [x86_64-linux-4.9.30-unRAID] (local build) Copyright (C) 2002-16, Bruce Allen, Christian Franke, www.smartmontools.org Short INQUIRY response, skip product id A mandatory SMART command failed: exiting. To continue, add one or more '-T permissive' options. I tried changing the SATA connections (both power and data) and unRAID was still showing issues, but at this point didn't try to mount or run a self-test. I now have the drive connected using the original data cable but with a new power cable, with the array stopped I was able to mount the drive and get a SMART report. On running the short self-test unRAID shows there's an error, however the SMART reports overall health is showing as PASSED with no error's logged which is a bit confusing. The test is though shown under the self-test log with a status of "Completed: read failure". unRAID highlights the issue as "Current pending sector": I have attached the full report as a txt file rather than include it in full on this post. So this re-purposed drive had successfully been pre-cleared, reformatted and added to the array, however it is now showing as faulty after a couple days of being online without any direct usage. Would it be reasonable to continue using this disk or would this be a must replace kind of deal? It is one of four re-purposed drives, would it be advisable to run extended SMART self-checks on the other drives? or some other kind of test? Many thanks WDC_WD30EFRX-68EUZN0_WD-WCC4N4SCEYTH-20170831-1250.txt Quote Link to comment
JonathanM Posted August 31, 2017 Share Posted August 31, 2017 22 minutes ago, Carrosive said: On moving the drives I installed a Supermicro AOC-SASLP-MV8 for the additional SATA connections This is a marvell based controller, there are known issues with them. Try disabling any virtualization options in BIOS, and make sure the firmware on the card and your motherboard's BIOS are up to date. Quote Link to comment
JorgeB Posted August 31, 2017 Share Posted August 31, 2017 You should also run an exte 49 minutes ago, Carrosive said: The test is though shown under the self-test log with a status of "Completed: read failure" Read failure means the test failed an the disk needs to be replaced, or if you prefer preclear it to see if the pending sector(s) are remapped. Quote Link to comment
Carrosive Posted August 31, 2017 Author Share Posted August 31, 2017 Thanks for the responses! @jonathanm Ah, I bought this card after seeing it in the UK recommended build on the hardware compatibility wiki page and being able to get it inexpensively. After reading your post I've come across other threads mentioning instability with the card which can be improved with disabling virtualisation as you say, however I would like to be able to run VM's on my server. The motherboard BIOS update is straight forward enough and I've managed to find instructions for updating the cards firmware. Since I've just started using unRAID, and as I can easily return the card and go without the additional hard drives for a while, I'm wondering if I should just bite the bullet and get something better to save me from headaches in the future. Something like the SilverStone SST-ECS02 (rebranded LSI SAS 9211-8i) seems like a relatively inexpensive alternative, I'm aware it will need flashed for IT mode. Any thoughts? @johnnie.black Tried to run an extended test however it failed shortly after starting, like it did with the short test, due to read failure. The drive was working fine and has gone unused so I'm leaning towards the issue being with my SATA card, I'll attempt to reassign it to the array when I get the card sorted out. Thanks for the info! Quote Link to comment
JorgeB Posted August 31, 2017 Share Posted August 31, 2017 Just now, Carrosive said: The drive was working fine and has gone unused so I'm leaning towards the issue being with my SATA card SMART test failure has nothing to do with the controller, disk has bad sector(s). Quote Link to comment
Carrosive Posted August 31, 2017 Author Share Posted August 31, 2017 @johnnie.black I think I may be confusing info from all the threads I've been looking at. Replacement would be ideal, do you know if this is something that would be covered under warranty? I think I'll do extended tests on the other cards to see what they're like. Thanks again Quote Link to comment
JorgeB Posted August 31, 2017 Share Posted August 31, 2017 Just now, Carrosive said: do you know if this is something that would be covered under warranty? Yes, SMART data shows a pending sector and SMART tests fail, that's more than enough to get an RMA. Quote Link to comment
Carrosive Posted August 31, 2017 Author Share Posted August 31, 2017 @johnnie.black Awesome, I'll see if the extended tests show any issues with the other drives and request RMA I believe it was a post of yours I saw in another thread from someone having an issue with their AOC-SASLP-MV8, recommending to get a LSI card instead. I'm considering getting a SilverStone SST-ECS02 (rebranded LSI SAS 9211-8i) to use instead as I'd prefer my server to be more stable and I'd like to keep virtualisation enabled, any thoughts? Quote Link to comment
JorgeB Posted August 31, 2017 Share Posted August 31, 2017 (edited) LSI are the recommended controllers for unRAID v6, although not every user has issues with the SASLP or SAS2LP, the number of users who do is significant, so they should be avoided whenever possible. Any LSI based on the SAS2008 or 2308 chipsets in IT mode will work great, e.g., 9201-8i, 9211-8i, 9207-8i and clones. Edited August 31, 2017 by johnnie.black Quote Link to comment
Carrosive Posted August 31, 2017 Author Share Posted August 31, 2017 Great - managed to find a reasonably priced Fujitsu D2607 (SAS2008) pre flashed in IT mode so I have that on the way. Thanks again! 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.