October 27, 201213 yr 2 days ago I installed sabnzbd and sickbeard on the cache drive, and assume I set it up right for the plugins to just use the cache, and my media share to not. Yesterday I installed simplefeatures. Last night the web interface for the array started becoming unstable and was very difficult to reach. I had to do an unclean powerdown and sit through a parity check. I started poking around this monring and found #1 the media share seems to be using the cache(I don't know why) and a ton of errors in the unmenu system log. Oct 27 08:56:10 Tower kernel: sd 1:0:0:0: [sdb] Unhandled error code (Errors) Oct 27 08:56:10 Tower kernel: sd 1:0:0:0: [sdb] Result: hostbyte=0x04 driverbyte=0x00 (System) Oct 27 08:56:10 Tower kernel: sd 1:0:0:0: [sdb] CDB: cdb[0]=0x28: 28 00 00 04 01 a0 00 00 08 00 (Drive related) over and over and over again. Can someone help me out?
October 27, 201213 yr 2 days ago I installed sabnzbd and sickbeard on the cache drive, and assume I set it up right for the plugins to just use the cache, and my media share to not. Yesterday I installed simplefeatures. Last night the web interface for the array started becoming unstable and was very difficult to reach. I had to do an unclean powerdown and sit through a parity check. I started poking around this monring and found #1 the media share seems to be using the cache(I don't know why) and a ton of errors in the unmenu system log. Oct 27 08:56:10 Tower kernel: sd 1:0:0:0: [sdb] Unhandled error code (Errors) Oct 27 08:56:10 Tower kernel: sd 1:0:0:0: [sdb] Result: hostbyte=0x04 driverbyte=0x00 (System) Oct 27 08:56:10 Tower kernel: sd 1:0:0:0: [sdb] CDB: cdb[0]=0x28: 28 00 00 04 01 a0 00 00 08 00 (Drive related) over and over and over again. Can someone help me out? We can't tell either, you did not include a syslog as an attachment, and you did not include a smartctl report on drive /dev/sdb. Both of those are needed for anyone to perform any analysis. Joe L.
October 27, 201213 yr Author I downloaded the syslog from unmenu and it says its 1.73 GB, is it me or is that insane? Plus much larger than able to attach right? Where do I find smartctl?
October 27, 201213 yr Author Zipping my syslog resulted in a file 2.2MB which is still larger than I can attach. What can I do to attach it? When I attempted to run smartctl it worked on my data drives, but not on the cache. The cache is an old 40GB SATA I, it's possible it doesn't have smart features right? As added info RAM cache is high - using about 5-6GB and when attempting to stop the array I get unmounting disks....retry unmounting disks... over and over.
October 27, 201213 yr Zipping my syslog resulted in a file 2.2MB which is still larger than I can attach. What can I do to attach it? When I attempted to run smartctl it worked on my data drives, but not on the cache. The cache is an old 40GB SATA I, it's possible it doesn't have smart features right? As added info RAM cache is high - using about 5-6GB and when attempting to stop the array I get unmounting disks....retry unmounting disks... over and over. You must stop all your add-on processes before the disks they are accessing can be un-mounted. All SATA disks will have SMART capabilities. Odds of your cache drive not having it are very slim. Post your syslog on a pastebin site or something similar. Or, cut the first 1/4 of of it, zip it, and post it. You'll probably run out of RAM soon enough... as the syslog is in memory. At that point the only thing you'll be able to do is a hard reboot. Joe L.
October 27, 201213 yr Author I'm having some trouble here, sorry. Tried to open the .txt and its so large I'm having trouble with it. Only program I have that will open it is Wordpad, and it's making it very unstable. Impossible to copy and paste. I guess I've never worked with a 2GB txt file before. I thought influencers plugins did not have to be stopped prior to stopping the array. I may have made the problem better or worse. The only way I could access is telnet, I found that there was an open python process running like crazy(61mil reads to the cache drive in a short time) so I killed it. Array stopped and was able to access the web GUI again, however now the cache drive shows unformatted.
October 27, 201213 yr Author Yes cache is sdb. I think I'm running it correctly and get the message: Smartctl: Device Read Identity Failed (not an ATA/ATAPI device)
October 27, 201213 yr Did you use this command? smartctl -a /dev/sdb Power down and check the connections to the drive. A SMART report is required.
October 28, 201213 yr Yes cache is sdb. I think I'm running it correctly and get the message: Smartctl: Device Read Identity Failed (not an ATA/ATAPI device) It may have died. (or the cable to it has come off) Does it show in ls -l /dev/disk/by-id
October 30, 201213 yr Author I've been out of town and will check this week and report back. Right now I have it disabled and have another concern that I want to work on first.
November 18, 201213 yr Author had to format the cache to remount it, finally was able to run a smart report. smart.txt
November 18, 201213 yr Unfortunately, that is not a smart report, but what the smartctl command will print when it is unable to establish communications with a drive. Did you invoke the smartctl with the "-d ata" argument or "-d sat" argument as suggested by smartctl in the output you attached? smartctl -a -d ata /dev/sdX or possibly: smartctl -a -d sat /dev/sdX If you did, then the disk is not responding. Unless the cable to the drive is bad (or not seated properly), or it has no power (or the power cable not seated properly), or its drive tray is not seated, or its disk controller port has died, the disk itself has probably died. Joe L.
November 18, 201213 yr Author That's strange, when I ran smartctl - a - A /dev/sdc I receive a report part of which says something like overall health failed drive likely to fail within 24 hours.
November 18, 201213 yr That's strange, when I ran smartctl - a - A /dev/sdc I receive a report part of which says something like overall health failed drive likely to fail within 24 hours. can you post the full output of smartctl -d ata -a /dev/sdc It should say the same, but give the failing attributes too.
November 18, 201213 yr There is the answer. ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE 1 Raw_Read_Error_Rate 0x000f 001 001 051 Pre-fail Always FAILING_NOW 25576 apparently, the disk cannot read what is on the platters... it is not sector specific. RMA the drive, it has died.
November 18, 201213 yr Author Thanks Joe L. It was an old drive, I'll just replace it. I know many reccomend using a size matched to parity for cache since you could use it as a quick swap, however if I'm running SB, SAB, and others on my cache wouldn't I lose those settings on all those apps if I ended up doing this? Is it worth buying a 3TB drive for cache to match parity.
November 18, 201213 yr Is it worth buying a 3TB drive for cache to match parity. Only if you can get it cheep. Otherwise, use what you have. The settings (and apps) can be copied off onto one of your data drives. and only copied back if the cache drive needs to be re-purposed.
Archived
This topic is now archived and is closed to further replies.