May 24, 201016 yr My server has been running for a LONG time, it has probably been 6 months since I rebooted it. Today all the shares suddenly became unavailable. I want to the server and I coud see the sharea but could not access any of them (though I could access each drive that is shared out... like drive1, drive2 would work but \\tower\data would not work). I rebooted the machine. When it came back up it would not start the array. It just says "Stopped. Configuration Valid." which is weird, because I would think it would say the configuration was invalid? Starting the array does not start it. Doing a Restore makes everything blue, then all the drives say "mounting". Once the mounting goes away I am back to an array that is stopped with a configuration valid message. Any ideas?
May 24, 201016 yr My server has been running for a LONG time, it has probably been 6 months since I rebooted it. Today all the shares suddenly became unavailable. I want to the server and I coud see the sharea but could not access any of them (though I could access each drive that is shared out... like drive1, drive2 would work but \\tower\data would not work). I rebooted the machine. When it came back up it would not start the array. It just says "Stopped. Configuration Valid." which is weird, because I would think it would say the configuration was invalid? Starting the array does not start it. Doing a Restore makes everything blue, then all the drives say "mounting". Once the mounting goes away I am back to an array that is stopped with a configuration valid message. Any ideas? You should have grabbed and posted a copy of the syslog. Pressing the button labeled as "Restore" is probably the worst thing you could have done. It sets a new initial configuration and invalidates any previous parity calculations. Please post a syslog. Instructions under troubleshooting in the wiki: http://lime-technology.com/wiki/index.php?title=Troubleshooting
May 24, 201016 yr Author I had a friend who is NOT colorblind look at my web page... It seems that the parity drive is red. I assume this means it is failed and I need to replace it. Do I have other options, like issue a format or rebuild to it? I will do some more searches on this board to see what my options are. If someone happens to reply to this thread, thank you for your time. On a side note... 30% of the population is colorblind, the least you could do is make the icons square or triangle or something.
May 24, 201016 yr I had a friend who is NOT colorblind look at my web page... It seems that the parity drive is red. I assume this means it is failed and I need to replace it. Do I have other options, like issue a format or rebuild to it?It might mean you pressed the button labeled as "Restore" but that is more of an orange color on my monitor. Or it might have failed. The ONLY WAY TO KNOW is for you to analyze your system log. Since most people do not know what to look for, most will follow the recommendation to capture and attach a copy of the system log to their request for help. It is the ONLY way anybody can tell what is happening. I will do some more searches on this board to see what my options are. If someone happens to reply to this thread, thank you for your time.No problem, just post a syslog, and do NOT press the button labeled as "restore" It has caused so many issues when pressed at the wrong time that it has been removed from the current 4.5.4 version of unRAID. Pressing it sets a new initial configuration based on the currently assigned and working disks as if you were setting up the server for the very first time. It has NOTHING to do with restoring data. On a side note... 30% of the population is colorblind, the least you could do is make the icons square or triangle or something. Excellent point. Send [email protected] an e-mail stating exactly that. It is a very good idea to make the error more noticeable, even for those not color-blind. Joe L.
May 24, 201016 yr Author Thank you for your responses... I have attached the syslog (I had installed UnMenu a while ago and forgot about it) Took me a bit to figure out how to get at that webpage versus the typical web page. From what I can tell my parrity array is invalid, which means (I would assume) that it is either dead or a complete power down is required. SHame too, this was my newest drive. I will probably upgrade to the latest version once I get this worked out.
May 24, 201016 yr On a side note... 30% of the population is colorblind, the least you could do is make the icons square or triangle or something. Excellent point. Send [email protected] an e-mail stating exactly that. It is a very good idea to make the error more noticeable, even for those not color-blind. Especially since red/green is the most common type of dichromacy, so 'DISK_OK' and 'DISK_DSBL' would both look identical (sort of grey, right?).
May 24, 201016 yr I'm seeing many many errors on disk "sdc" in your syslog: May 24 18:00:30 MediaServer kernel: sd 2:0:0:0: [sdc] 976773168 512-byte hardware sectors (500108 MB) May 24 18:00:30 MediaServer kernel: sd 2:0:0:0: [sdc] Write Protect is off May 24 18:00:30 MediaServer kernel: sd 2:0:0:0: [sdc] Mode Sense: 00 3a 00 00 May 24 18:00:30 MediaServer kernel: sd 2:0:0:0: [sdc] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA May 24 18:00:30 MediaServer kernel: ata2.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0 May 24 18:00:30 MediaServer kernel: ata2.00: irq_stat 0x40000001 May 24 18:00:30 MediaServer kernel: ata2.00: cmd c8/00:08:00:00:00/00:00:00:00:00/e0 tag 0 dma 4096 in May 24 18:00:30 MediaServer kernel: res 51/01:00:00:00:00/85:00:00:00:00/e0 Emask 0x1 (device error) May 24 18:00:30 MediaServer kernel: ata2.00: status: { DRDY ERR } May 24 18:00:30 MediaServer kernel: ata2.00: configured for UDMA/133 May 24 18:00:30 MediaServer kernel: ata2: EH complete May 24 18:00:30 MediaServer kernel: sd 2:0:0:0: [sdc] 976773168 512-byte hardware sectors (500108 MB) May 24 18:00:30 MediaServer kernel: sd 2:0:0:0: [sdc] Write Protect is off May 24 18:00:30 MediaServer kernel: sd 2:0:0:0: [sdc] Mode Sense: 00 3a 00 00 May 24 18:00:30 MediaServer kernel: sd 2:0:0:0: [sdc] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA May 24 18:00:30 MediaServer kernel: ata2.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0 May 24 18:00:30 MediaServer kernel: ata2.00: irq_stat 0x40000001 May 24 18:00:30 MediaServer kernel: ata2.00: cmd c8/00:08:00:00:00/00:00:00:00:00/e0 tag 0 dma 4096 in May 24 18:00:30 MediaServer kernel: res 51/01:00:00:00:00/85:00:00:00:00/e0 Emask 0x1 (device error) May 24 18:00:30 MediaServer kernel: ata2.00: status: { DRDY ERR } May 24 18:00:30 MediaServer kernel: ata2.00: configured for UDMA/133 May 24 18:00:30 MediaServer kernel: ata2: EH complete May 24 18:00:30 MediaServer kernel: sd 2:0:0:0: [sdc] 976773168 512-byte hardware sectors (500108 MB) May 24 18:00:30 MediaServer kernel: sd 2:0:0:0: [sdc] Write Protect is off May 24 18:00:30 MediaServer kernel: sd 2:0:0:0: [sdc] Mode Sense: 00 3a 00 00 May 24 18:00:30 MediaServer kernel: sd 2:0:0:0: [sdc] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA May 24 18:00:30 MediaServer kernel: ata2.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0 May 24 18:00:30 MediaServer kernel: ata2.00: irq_stat 0x40000001 May 24 18:00:30 MediaServer kernel: ata2.00: cmd c8/00:08:00:00:00/00:00:00:00:00/e0 tag 0 dma 4096 in May 24 18:00:30 MediaServer kernel: res 51/01:00:00:00:00/85:00:00:00:00/e0 Emask 0x1 (device error) May 24 18:00:30 MediaServer kernel: ata2.00: status: { DRDY ERR } May 24 18:00:30 MediaServer kernel: ata2.00: configured for UDMA/133 May 24 18:00:30 MediaServer kernel: ata2: EH complete I suspect that disk has failed, or the cable to it has become loose (either power or SATA cable) The only way to know is to run a "smart" report on /dev/sdc (You can request the smart status from the disk management page in unMENU) You can also request an "hdparm" report from the same place in unMENU. I'm seeing all kinds of indications the drive is not responding, so power down and check the cabling first. Try to not disturb the cabling on any other drive. You do not want to complicate your life with 2 failed drives. You have vastly complicated the restoration of /dev/sdc by pressing "restore" and invalidating parity. We need to get the array to think disk2 is the defective one, and that parity is correct, and then go from there. right now, it thinks parityis invalid, and that disk2 is fine, but it cannot even start the array to the point where it can mark disk2 as failed. Since the array is not started, I think you can issue a command to force disk2 to not be valid and attempt to come up online. 1.First we need to press the "restore" button once more to force a new initial configuration. then, BEFORE you attempt to start the array 2. log onto the system console (or via telnet) and type: /root/mcmd set invalidslot 2 It should respond OK. 3. Then, go back to the management web-page and press "Start" (You might need to check the checkbox under it to enable it) When the array starts, it will want to re-construct the failed disk 2's contents onto what it thinks is the replacement disk2. (pressing restore cause it to forget the disk's serial number) Let it begin this process. If it fails, then we'll have another clue that disk2 has failed. You can then stop the array, un-assign it, and re-start the array (keeping your fingers as far away from the button labeled as "restore" as possible. The button labeled as restore has gotten so many people like yourself into trouble it has been removed from the most current version 4.5.4 of unRAID. It has been replaced with an "initconfig" command that sets a new initial configuration. Joe L.
May 25, 201016 yr Author So I took my machine apart last night. I am certain the hard drive is bad. I went out today and got a slightly larger drive to use as a new parity disk. do you still want me to do the above steps or should I go through the normal steps of a failed parity drive?
May 25, 201016 yr The failed drive with all the errors in your syslog is /dev/sdc ( disk2 ), not your parity drive. May 24 18:12:59 MediaServer kernel: md: import disk2: [8,32] (sdc) WDC WD5000AAJS-00TKA0 WD-WCAPW5453654 offset: 63 size: 488386552 You can do as you said, but you may lose whatever data was on disk2. If you follow the steps I outlined, it might be possible to recover what was on disk2. It is entirely up to you how to proceed. Introducing a "slightly" larger drive right now will complicate things, since it can only be put in the parity slot. I'd not introduce any new drive at this time. Did you check the cabling on disk2 yet? Did you run a smart report on it yet? smartctl -d ata -a /dev/sdc Joe L.
May 25, 201016 yr Author I did check the cabling and it was good. I checked with the bios and at the bios level it could tell there was a drive but could not tell the number of heads/cylinders etc. That drive is toast. I wanted to be safe until I got new drives so I powered down the system and left it off. When I get home tonight I will follow the steps you posted above. I do have a roughly same size drive I can put in (I got a 1tb to replace the parity and a 500gig for additional space). I will follow the steps above before adding anything else. I would really hate to lose any data. Any other advice? And thanks again for all the help, I am used to hardware based raids which act a bit differently (keeping raid config data both on the drives and on the controller).
May 26, 201016 yr Author So I followed the steps you had me do. They seemed to work. I have attached the new syslog. The array can be brought up but disk 2 says unassigned, so it didn't start a rebuild. Have I lost all the data on that drive... that could be REAL bad. new_syslog.zip
May 26, 201016 yr OK, Now we need to deal with this on disk2 May 26 03:25:01 MediaServer kernel: ReiserFS: md2: found reiserfs format "3.6" with standard journal May 26 03:25:01 MediaServer kernel: ReiserFS: md2: warning: Filesystem on md2 cannot be mounted because it is bigger than the device May 26 03:25:01 MediaServer kernel: ReiserFS: md2: warning: You may need to run fsck or increase size of your LVM partition May 26 03:25:01 MediaServer kernel: ReiserFS: md2: warning: Or may be you forgot to reboot after fdisk when it told you to and May 26 03:25:01 MediaServer kernel: md1: running, size: 488386552 blocks May 26 03:25:01 MediaServer kernel: md2: running, size: 0 blocks May 26 03:25:01 MediaServer kernel: md3: running, size: 312571192 blocks May 26 03:25:01 MediaServer kernel: md4: running, size: 312571192 blocks May 26 03:25:01 MediaServer kernel: md5: running, size: 312571192 blocks May 26 03:25:01 MediaServer kernel: md6: running, size: 312571192 blocks May 26 03:25:01 MediaServer kernel: md7: running, size: 488386552 blocks May 26 03:25:01 MediaServer kernel: md8: running, size: 488386552 blocks The first line is encouraging... since it seems to indicate the virtual file-system reconstructed by parity and the other drives is present, but the file-system partition is not sized properly. I'm thinking that the disk reported a size of 0 (or made no report at all, as you've seen in the bios listing). Or, did you swap around disks in your earlier attempt to fix things? Did you specifically un-asssign disk2 on your devices page? Or did you not get to that step in the series of steps I described earlier? Do you have any spare ports on your disk controllers? (Where we could attach one of your spare drives, but keep it out of the array for a short time, but still use it in the recovery effort. In other words, not assign it on the "Devices" page, but still access it through command line commands.) It appears as if unRAID is assigning the /dev/sdc to a different disk now, so the above /dev/sdc is probably not the drive that has failed. (In fact, it looks to be that /dev/sdc is your parity drive and disk2 was not detected and is "missing") May 26 03:40:29 MediaServer kernel: md: import disk0: [8,32] (sdc) WDC WD7500AACS-00D6B1 WD-WCAU4A986180 offset: 63 size: 732574552 If you have a spare disk that is as big (or bigger) than your original disk2, but not bigger than your parity disk, you can install it and assign it as disk2. We can then see if it will allow the re-construction to begin. Joe L.
May 26, 201016 yr Author At your behest I swapped no disks. I followed your earlier instructions but when I got to the part about re-assigning the disk2 back to itself I stopped. That drive is really in bad shape and thought I would get more advice before proceeding. I have attached some screen caps to help things out. I did add a new disk to the system but left it unassigned as I wanted more advice on the best way to go about this, it is the same size as the one that died (500 gigs) but is smaller than the parity drive (720 gigs). screencap.zip
May 26, 201016 yr Author Here is the other screen cap. I will assign the new unassigned disk as disk 2 and hope it rebuilds. Did I say thanks for the help? screencaps.zip
May 26, 201016 yr Author Yippppppeeeeeeee It looks like replacing the drive caused a rebuild to start. Once this rebuild is done, I expect to 1. upgrade to the latest version. 2. remove the old dead drive and see if it is still under mfgr warranty 3. add a new 1tb drive and follow the procedure for making that the parity drive. thanks for all the help.
May 26, 201016 yr At your behest I swapped no disks. I followed your earlier instructions but when I got to the part about re-assigning the disk2 back to itself I stopped. That drive is really in bad shape and thought I would get more advice before proceeding. I have attached some screen caps to help things out. I did add a new disk to the system but left it unassigned as I wanted more advice on the best way to go about this, it is the same size as the one that died (500 gigs) but is smaller than the parity drive (720 gigs). Do you know the "device" name of the drive you added? (I guess I can try to deduce it from the last syslog, but it would save me a few minutes) Your screen caps did not get attached. Perhaps you can try again with them. If you did not un-assign disk2, this would be the time to do it on the devices page. That should take the defective disk out of the picture. See if the array will then "Start" Again, don't do anything but press "Start" at this point. I'd like to see the screen-shot before proceeding so I can see what you are seeing. Joe L.
May 26, 201016 yr You are a bit ahead of me, but great news that a rebuild is in progress. Can you see the files on it from your file-explorer in windows? Post a syslog, now, before it completes the re-construction. It will give me an idea of what is happening. Joe L.
May 26, 201016 yr Looking at the old screen shot, it seems to indicate that disk2 is unformatted. Does it still say that? DO NOT PRESS THE FORMAT BUTTON. DO NOT PRESS THE FORMAT BUTTON. Let me repeat DO NOT PRESS THE FORMAT BUTTON. DO NOT PRESS THE FORMAT BUTTON. To unRAID, it just indicates the drive could not be mounted. (and any drive it cannot mount it thinks is unformatted) Just post a syslog. We may need to do a file-system repair for it to mount, but I won't know until you answer some of the previous questions. Joe L. Did I mention that you SHOULD NOT PRESS THE RESTORE BUTTTON EITHER (in case you forgot) ;D
May 26, 201016 yr Author So at your behest... I pushed the format button (4 times like you said). then I pushed the Restore button, just to make sure I was doing it right. Just kidding. attached is a screen cap of the progress. Syslog coming soon.
May 26, 201016 yr So at your behest... I pushed the format button (4 times like you said). then I pushed the Restore button, just to make sure I was doing it right. Just kidding. attached is a screen cap of the progress. Syslog coming soon. That screen cap looks wonderful.... I can see that disk2 did get mounted, and that the re-construction is under way. And there is no "Format" button, since there are no disks it thinks are unformatted. (So you can't press it) Now... Do not allow the server to lose power until the disk is re-constructed. Joe L. (I realize it is not as easy to keep the power coming into your home, but we want the process to complete without interruption)
May 26, 201016 yr Based on the latest syslog, you are not out of the woods just yet. It shows the file-system on disk2 needs repair. May 26 15:59:49 MediaServer kernel: ReiserFS: warning: is_tree_node: node level 0 does not match to the expected one 3 May 26 15:59:49 MediaServer kernel: ReiserFS: md2: warning: vs-5150: search_by_key: invalid format found in block 35934207. Fsck? May 26 15:59:49 MediaServer kernel: ReiserFS: md2: warning: vs-13070: reiserfs_read_locked_inode: i/o failure occurred trying to find stat data of [900 958 0x0 SD] May 26 15:59:49 MediaServer kernel: ReiserFS: warning: is_tree_node: node level 0 does not match to the expected one 3 May 26 15:59:49 MediaServer kernel: ReiserFS: md2: warning: vs-5150: search_by_key: invalid format found in block 35934207. Fsck? May 26 15:59:49 MediaServer kernel: ReiserFS: md2: warning: vs-13070: reiserfs_read_locked_inode: i/o failure occurred trying to find stat data of [900 909 0x0 SD] Let the re-construction complete. At that point, we'll perform the file-system repair. We can't really do that until the re-construction is finished (or rather, I prefer to wait until it is complete) Joe L.
May 26, 201016 yr You'll need to follow the procedure as described here in the wiki: http://lime-technology.com/wiki/index.php?title=Check_Disk_Filesystems You'll be operating on /dev/md2 so use the appropriate commands for that disk. I need to mow the lawn, so I'll get back to this thread in a few hours. By then, the re-construction might be close to finishing. Then, I have a dance lesson and I'll check in again after it. (I'm not completely a geek, I have some social skills, believe it or not ) Joe L. Ballroom dancing is one of our hobbies. I look almost exactly like the dancers you see on "Dancing with the Stars" except: I keep my clothes on, don't do lifts, don't throw my dance partners across the floor, and I don't wax my chest. Now, if I could only dance as well ...
May 26, 201016 yr Author I ran the check... It said I must run a --rebuild-tree (see pic) the instructions you posted for me to read said NOT to run it until I talked to poeple on the forum.
Archived
This topic is now archived and is closed to further replies.