July 7, 201510 yr Looking for some assistance on my very first red ball (I’ve been running unRAID since ~2009). I was planning to upgrade my server from v5.0.6 to v6. Of course, before doing so, I decided to run a parity check. It has been about 90 days since my last parity check. Anyhow, I awoke this morning to find that my disk15 red-balled during the parity check. I’ve read through various posts on the forum and some of you seem to be much better than I am at understand unRAID logs and determining the problem. Below, I’ve included links to both my recent syslog and the output of smartctl that I ran on disk15 this evening. The smartctl doesn’t seem to indicate any errors. However, there are thousands of read/write errors in the syslog. Before I proceed with any corrections to the issue I was hoping to get some feedback on my logs. I see the following in the log: Jul 6 08:47:32 media kernel: ata21.00: exception Emask 0x0 SAct 0x7ff87e3d SErr 0x0 action 0x6 frozen I also see this message appear a lot: Jul 6 08:47:32 media kernel: ata21.00: failed command: READ FPDMA QUEUED Jul 6 08:47:32 media kernel: ata21.00: cmd 60/88:00:98:27:9b/00:00:b6:00:00/40 tag 0 ncq 69632 in Jul 6 08:47:32 media kernel: res 40/00:64:50:38:57/00:00:38:00:00/40 Emask 0x4 (timeout) Jul 6 08:47:32 media kernel: ata21.00: status: { DRDY } As a side note, I’m using a Norco 4224 and haven’t opened the case in a VERY long time. Of course, that doesn’t necessarily mean that it may not be something as simple as a loose cable. Currently, I can access my disk via /mnt/disk15. I can see my folder structure but that doesn’t necessarily mean that all the data is intact. Anyhow, I would seriously appreciate any assistance the experts can provide with my logs. -- James syslog (abridged, but still too large to attach): https://drive.google.com/file/d/0B-vJmhFsmL8idGRHSHBfY254Z1k/view?usp=sharing smartctl output: https://drive.google.com/file/d/0B-vJmhFsmL8ibDNCbU5yVXBCQjg/view?usp=sharing
July 7, 201510 yr unRAID redballs a disk because of a write failure. This means unRAID has disabled the disk and is not actually using it now. When you look at the contents of the disk, unRAID is actually simulating the drive by calculating its data from reading all the other drives plus parity. It is not obvious from the smart that the drive is bad so it might be a problem with a connection. Nevertheless, the drive will have to be rebuilt from parity. Do you have a spare?
July 7, 201510 yr Author unRAID redballs a disk because of a write failure. Interesting, I do see a lot of write failures in my log file. Does default unRAID parity check attempt to write to my disk? When you look at the contents of the disk, unRAID is actually simulating the drive by calculating its data from reading all the other drives plus parity. Duh .. of course, I don't know why I didn't recognize this. I'm a little bit ashamed of myself It is not obvious from the smart that the drive is bad so it might be a problem with a connection. Going to check connections and reseat drives right now. Is there a way to re-run my parity check without having to rebuild disk15 first? If there was simply a communication problem during my previous parity check, is there a possibility that data on disk15 is still intact and correct? Do you have a spare? I do not, but I will in 3-5 business days. I just ordered a new drive from newegg.
July 7, 201510 yr unRAID redballs a disk because of a write failure. Interesting, I do see a lot of write failures in my log file. Does default unRAID parity check attempt to write to my disk? Not unless a read fails, in which case parity is used to calculate what should be there, and unraid attempts to write the parity generated value. If the write fails, the drive is red balled.
July 7, 201510 yr Author unRAID redballs a disk because of a write failure. Interesting, I do see a lot of write failures in my log file. Does default unRAID parity check attempt to write to my disk? Not unless a read fails, in which case parity is used to calculate what should be there, and unraid attempts to write the parity generated value. If the write fails, the drive is red balled. Thanks, that's the answer I was expecting to hear. Which means that, by default, an unRAID parity check is a "correcting" parity check. Is that correct?
July 7, 201510 yr unRAID redballs a disk because of a write failure. Interesting, I do see a lot of write failures in my log file. Does default unRAID parity check attempt to write to my disk? Not unless a read fails, in which case parity is used to calculate what should be there, and unraid attempts to write the parity generated value. If the write fails, the drive is red balled. Thanks, that's the answer I was expecting to hear. Which means that, by default, an unRAID parity check is a "correcting" parity check. Is that correct? It does default to a correcting check, however, even an uncorrecting check reads the drives, which will trigger the same mechanism that writes to the drive that failed the read, for better or worse. My personal preference would be a way to select a mode that was completely read only on all drives, and just report errors instead of taking any action. The difference between correcting and non-correcting checks is purely whether or not a successfully read but incorrect parity value is written to the parity disk. Any read errors are treated separately.
July 7, 201510 yr Author It is not obvious from the smart that the drive is bad so it might be a problem with a connection. I just reseated my drive and the SAS cable connecting my Norco 4224 backplane to my RAID controller card. How can I tell if this resolves the issue? Should I attempt to rebuild the drive onto itself? If so, do I need to pre-clear the drive first?
July 7, 201510 yr It is not obvious from the smart that the drive is bad so it might be a problem with a connection. I just reseated my drive and the SAS cable connecting my Norco 4224 backplane to my RAID controller card. How can I tell if this resolves the issue? Should I attempt to rebuild the drive onto itself? If so, do I need to pre-clear the drive first? No need to preclear. Do you know how to make unRAID rebuild onto itself?
July 7, 201510 yr Author It is not obvious from the smart that the drive is bad so it might be a problem with a connection. I just reseated my drive and the SAS cable connecting my Norco 4224 backplane to my RAID controller card. How can I tell if this resolves the issue? Should I attempt to rebuild the drive onto itself? If so, do I need to pre-clear the drive first? No need to preclear. Do you know how to make unRAID rebuild onto itself? I do not ... I have never done it before. However, I assume it is in the online unRAIDI manual. Is attempting to rebuild this drive onto itself a really bad idea?
July 7, 201510 yr It is not obvious from the smart that the drive is bad so it might be a problem with a connection. I just reseated my drive and the SAS cable connecting my Norco 4224 backplane to my RAID controller card. How can I tell if this resolves the issue? Should I attempt to rebuild the drive onto itself? If so, do I need to pre-clear the drive first? No need to preclear. Do you know how to make unRAID rebuild onto itself? I do not ... I have never done it before. However, I assume it is in the online unRAIDI manual. Is attempting to rebuild this drive onto itself a really bad idea? The rebuild will not affect parity or any of your other drives, and since the disk itself looks OK, it is probably OK to rebuild the drive onto itself. Stop the array. Unassign the disk. Start the array. unRAID will show this drive is missing. Stop the array. Assign the disk back to the same slot. Start the array. unRAID will rebuild.
July 7, 201510 yr Author Stop the array. Unassign the disk. Start the array. unRAID will show this drive is missing. Stop the array. Assign the disk back to the same slot. Start the array. unRAID will rebuild. Awesome ... thanks trurl ... I was just looking through the manual for this info. I will test out a rebuild and see what happens.
July 7, 201510 yr Author I have successfully rebuilt the drive onto itself from parity. I'm running another parity check now to be sure nothing strange is happening.
July 7, 201510 yr Author :( Came home from work to see my parity check caused a red-ball on disk15 again. The main unRAID page shows that there were 2048 errors. This is the same number of errors that I saw when it red-balled yesterday. I have a couple of new 4TB disks on the way from newegg. I guess will try to rebuild disk15 onto one of the new drives when they arrive. I won't have the new disks until Thursday. Then I have to spend a day pre-clearing the replacement drive. Then rebuild disk15. And then try parity check again. Ugh ... it is going to be a long wait.
July 7, 201510 yr Author Trying to run smartctl on disk15 produces the following error: root@media:~# smartctl --all /dev/sdp smartctl 6.2 2013-07-26 r3841 [i686-linux-3.9.11p-unRAID] (local build) Copyright (C) 2002-13, Bruce Allen, Christian Franke, www.smartmontools.org === START OF INFORMATION SECTION === Vendor: /8:0:6:0 Product: >> Terminate command early due to bad response to IEC mode page A mandatory SMART command failed: exiting. To continue, add one or more '-T permissive' options.
July 12, 201510 yr Author Final update on my disk15 issue. I replaced the disk with a new (and larger disk), rebuilt the data successfully, and ran a parity check. Everything seems to be running normally now. My original disk15 HDD has been decommissioned since it appears to have some problems. Now I can finally upgrade to v6
Archived
This topic is now archived and is closed to further replies.