June 24, 201214 yr after a reboot i am having issues with read only file systems. i'm not sure how to view the output of reiserfsck. it says 'will log to stdout' - how do i view this? presumably i'll need to run a --rebuildtree ?? HELP!? syslog below: Jun 24 22:03:13 Tub avahi-daemon[11232]: Service group file /services/smb.service changed, reloading. Jun 24 22:03:13 Tub emhttp: shcmd (56): /usr/local/sbin/emhttp_event svcs_restarted Jun 24 22:03:13 Tub emhttp_event: svcs_restarted Jun 24 22:03:13 Tub emhttp: unclean shutdown detected, checking parity... Jun 24 22:03:13 Tub kernel: mdcmd (38): check CORRECT Jun 24 22:03:13 Tub kernel: md: recovery thread woken up ... Jun 24 22:03:13 Tub kernel: md: recovery thread syncing parity disk ... Jun 24 22:03:13 Tub kernel: md: using 1536k window, over a total of 2930266532 blocks. Jun 24 22:03:14 Tub avahi-daemon[11232]: Service "Tub-SMB" (/services/smb.service) successfully established. Jun 24 22:05:21 Tub shfs/user: shfs_create: open: /mnt/disk1/Media/downloads/Movies/._.DS_Store (5) Input/output error Jun 24 22:05:21 Tub kernel: REISERFS error (device md1): reiserfs-2025 reiserfs_cache_bitmap_metadata: bitmap block 87195648 is corrupted: first bit must be 1 Jun 24 22:05:21 Tub kernel: REISERFS (device md1): Remounting filesystem read-only Jun 24 22:05:26 Tub shfs/user: shfs_open: open: /mnt/disk1/Media/.DS_Store (30) Read-only file system Jun 24 22:05:33 Tub last message repeated 3 times Jun 24 22:05:33 Tub shfs/user: shfs_rename: rename: /mnt/disk1/Media/downloads/Movies/Pixar Short Film Collection 1080p BluRay x264 HD4U /mnt/disk1/Media/Movies/Pixar Short Film Collection 1080p BluRay x264 HD4U (30) Read-only file system Jun 24 22:05:36 Tub shfs/user: shfs_rename: rename: /mnt/disk1/Media/downloads/Movies/Pixar Short Film Collection 1080p BluRay x264 HD4U /mnt/disk1/Media/Movies/Pixar Short Film Collection 1080p BluRay x264 HD4U (30) Read-only file system Jun 24 22:05:46 Tub shfs/user: shfs_open: open: /mnt/disk1/Media/.DS_Store (30) Read-only file system Jun 24 22:05:51 Tub last message repeated 3 times Jun 24 22:05:51 Tub shfs/user: shfs_rename: rename: /mnt/disk1/Media/downloads/Movies/Pixar Short Film Collection 1080p BluRay x264 HD4U /mnt/disk1/Media/Movies/Pixar Short Film Collection 1080p BluRay x264 HD4U (30) Read-only file system Jun 24 22:05:55 Tub shfs/user: shfs_rename: rename: /mnt/disk1/Media/downloads/Movies/Pixar Short Film Collection 1080p BluRay x264 HD4U /mnt/disk1/Media/Movies/Pixar Short Film Collection 1080p BluRay x264 HD4U (30) Read-only file system Jun 24 22:08:08 Tub shfs/user: shfs_mkdir: mkdir: /mnt/disk1/Media/downloads/TV/Silk.2x01.720p HDTV x264-FoV (30) Read-only file system
June 24, 201214 yr Author i'd like to add that i only built the server this weekend and there is less than 10gb of data on there. i could very easily just wipe and start again. i didn't preclear the drives as i thought it would take too long. could this be the cause of such issues? the drives are sized as follows: 3x 3tb 2x 1tb 1x 1.5tb a few other ancillary drives under 500gb each
June 24, 201214 yr i'd like to add that i only built the server this weekend and there is less than 10gb of data on there. i could very easily just wipe and start again. i didn't preclear the drives as i thought it would take too long. could this be the cause of such issues? We have no idea. You did not attach a syslog, you did not even bother to tell us which version of unRAID you are running. the drives are sized as follows: 3x 3tb 2x 1tb 1x 1.5tb a few other ancillary drives under 500gb each About all we know is it probably is a licensed version, and probably one of the 5.X series. Yes, drive errors can cause all kinds of the issues you are experiencing. The preclearing process tends to find these early in the drive's life. Experience is that roughly 1 disk out of 5 have issues. You have 6 disks... If you want to get some experience with recovery procedures, I strongly suggest you follow the process in the wiki to fix your file-system errors. (You have nothing to lose now but time.) Then... think about starting fresh. http://lime-technology.com/wiki/index.php/Check_Disk_Filesystems Joe L.
June 24, 201214 yr Author sorry for not bothering. i'm on v5 rc5, with a pro license and i put in the first post the relevant part of the syslog - did you miss that? i just feel a little stuck. i've been going round in circles for hours and i'm very frustrated at this stage - only to come across a new error as soon as i've fixed the last one. my main question is how do i view the output of the reiserfsck it mentions stdout but i can't google successfully how to view this... so, what's the quickest way to get things working properly and the most thoroughly please? it can be a complete wipe, that's ok...
June 24, 201214 yr so, what's the quickest way to get things working properly and the most thoroughly please? it can be a complete wipe, that's ok... If you are serious that you don't care about any of your existing drive contents, then the best course of action is at least 1 preclear pass on every drive to verify its condition. It's not the quickest, but it's the most thorough. Without a full preclear, you really don't know if the drive will function properly. Since you only posted the portion of the syslog that you thought was relevant, there was no way to really analyze what's going on. Please attach a zipped copy of the full syslog, and smart test results on all the drives if you really want others to help you diagnose the situation.
June 24, 201214 yr Author thanks for your help. i'm going to start afresh and ditch this install but need to install screen to start the 6 drives off preclearing simultaneously. i found the wiki for how to it on v4, how do i do it on v5?
June 24, 201214 yr thanks for your help. i'm going to start afresh and ditch this install but need to install screen to start the 6 drives off. i found the wiki for how to it on v4, how do i do it on v5? Honestly, I'm still running 4.7, but I'm pretty sure unmenu and the screen version it lets you install should work fine on the 5 series.
June 24, 201214 yr Correct. Just start with a clean install of v5 and use screen per the config tutorial.
June 24, 201214 yr exactly. If you were an expert, you would know exactly what was relevant in the syslog. Since you are not, the snippit you posted was nowhere near sufficient. As others have already stated, you now must look at the reasons for the file-system corruption. Did you perform a memory test? or was that skipped too? (bad memory could easily cause corruption) Did one or more of your disks develop or have bad sectors? They too could cause file-system corruption. Next steps... A smartctl report on each of your drives, looking for sectors pending re-allocation, or those already re-allocated, an atached full zipped syslog, and a memory test, preferably overnight. (with no memory errors) If you do not determine why your file systems became corrupted, they will just continue to have issues and become corrupted again and again. If you have enough RAM you can perform all the preclear disks in parallel. If not, use the -r -w, and -b options to limit memory used. You do not need to use screen, you can use the 6 virtual consoles (assuming you have a monitor connected) or you can easily install screen. (It installs the same in any version of unRAID)
June 24, 201214 yr Author joe - i know you're here to help and some of your reply is infact very helpful, but i might suggest you're being a touch rude to me. i'm a COMPLETE novice here, as you say not an expert, and didn't realise the entire syslog was relevant (was just trying to make things easier). didn't know about the multiple console windows (alt f1 etc) till just now, thanks. how do i list each disk to get the name /dev/sda? etc...
June 24, 201214 yr Author answered a question by myself. eventually found the right forum post. if anyone searches here the easiest way to list drives is preclear_disk.sh -l i've also added my rig to my sig to make further troubleshooting easier for you experts.
June 24, 201214 yr joe - i know you're here to help and some of your reply is infact very helpful, but i might suggest you're being a touch rude to me. i'm a COMPLETE novice here, as you say not an expert, and didn't realise the entire syslog was relevant (was just trying to make things easier). didn't know about the multiple console windows (alt f1 etc) till just now, thanks. how do i list each disk to get the name /dev/sda? etc... I apologize if I came across as rude. It is not my intent. You however, came across as thinking all the advice about knowing the condition of your hardware before trusting your array to data as a waste of your time. (instead, you wish us to invest our time.. and many of us do) Most of what the unRAID community has learned is in the wiki. http://lime-technology.com/wiki/index.php/UnRAID_Wiki Truthfully, the user-contributed section is where you'll find the most up to date information. It is a great place to broaden your knowledge of unRAID and how to manage a server. I have more experience than many, and I still learn every single day. The multiple consoles are a real blessing, with the only disadvantage of it being more difficult to capture output of commands to post here. I've taken photos of the screen when that occurred, when I was reporting a bug, back in the early days of unRAID and all I had as a clue was the output on the system console. If we had the entire syslog (especially if the errors had occurred recently,) we might have seen the issue causing the corruption, and offered better advice. Your excerpt shows that the file-system on disk1 is corrupted. We do not know which disk that is to guide you, since the part of the syslog where the disks are inventoried is missing. All I could do is point you to the wiki. I could not give you any specific commands. It is as if you said "my car is making a funny noise... what part do I need to buy to fix it?" without telling the mechanic the make/model, or any details, they must take a test drive themselves before any educated guess can be offered. Again, this is a perfect time in a no stress situation to use the reiserfsck procedure as described in the wiki for you to learn. (It is hard to force a corruption in the file-system. You have some ready-made for you to practice on, with data that can easily be re-loaded if everything fails) If you have time this evening, test your memory. Let it run overnight while you sleep. Expect zero errors. To list all your drives, type: ls -l /dev/disk/by-id The device names with the trailing "1" are the devices to access specific partitions. For the smartctl command, you use the three character names /dev/sda, /dev/sdb, /dev/sdc, etc... smartctl -a /dev/sda smartctl -a /dev/sdb etc... The command you found preclear_disk.sh -l will only list those disks not yet assigned to the unRAID array. Joe L.
June 24, 201214 yr Author dude, we've all misinterpreted a written communication before now, maybe this is just me being tired a cranky! so don't worry - i got what i needed from you despite the confusion! anyway, current state of play is this... i'm preclearing 6 of my 7 drives currently. 8gb ram isn't likely to run into problems there is it? i also managed to get unmenu installed from google code. just my searches came together eventually! i'm going on a business trip for 4 days and was hoping (foolishly perhaps) to get the system up and running in a weekend... ah well, it's going to spend the next 40-50 hours preclearing the drives ready for my return on thursday when i will have lots of lovely questions about set ups of sabnzbd and such when using a cache drive. until then, thanks for all your help (it was helpful).
June 24, 201214 yr You do not need to wait for the preclear process to finish to supply the smart reports.. In fact, the initial reports created by the preclear script on those 6 drives are in /boot/preclear_reports on the flash drive (accessible as \\IP-ADDRESS\flash\preclear_reports from your finder... since it looks like you have a Mac) If you attach them to your next post, we can look at their beginning health. Joe L.
June 24, 201214 yr this is what i'm seeing, no reports.... yes, on my mac at the moment. sorry for the mis-information. I just checked... they are not copied there until later. For now, since you have unMENU, you can get them all from there on the disk-management page or from the myMain page where there are "sm" links for each drive. To get all the pages in unMENU working you'll need to add a hosts file entry on your Mac, or add a few entries in the unmenu.conf file to force it to use IP addresses on the sub-pages as described here: http://lime-technology.com/forum/index.php?topic=5568.msg51839#msg51839 Joe L.
June 24, 201214 yr Author taken the sm links from mymain and found that the 320 drive is fekked, or nearly! not surprising it came out of a sky+ box a while back. anyway, reports attached for any of the others i have missed anything useful for...? smreports.zip
June 25, 201214 yr (disk1 currently /dev/sda ) Device Model: ST1000DL002-9TT153 Serial Number: W1V01MBV Looks great. Only point of interest is 192 Power-Off_Retract_Count 0x0032 100 100 000 Old_age Always - 153 which indicates the disk heads were retracted in a power loss prior to a proper shutdown 153 times. (disk2 currently /dev/sdb ) Device Model: Hitachi HDS5C3030ALA630 Serial Number: MJ1311YNG4GE8A nothing of interest except again, disk heads being retracted in a power loss situation. (disk3 currently /dev/sdd ) Device Model: ST31500341AS Serial Number: 9VS0YEG8 5 Reallocated_Sector_Ct 0x0033 100 100 036 Pre-fail Always - 3 3 reallocated sectors in 25000 hours operation. Not too bad actually if not constantly getting new reallocated sectors. If these were recent, they could easily be the cause of the file system corruption, especially if you added the parity disk last. (after you formatted and loaded the other disks) (disk4 /dev/sdf ) Device Model: WDC WD3200AVVS-63L2B0 Serial Number: WD-WCAV12094972 looks fine as far as re-alloocated sectors... the error it shows seems to be where it was trying to write to its smart log. I would not be too concerned about it so far. (disk5 /dev/sdg ) Device Model: ST31000340AS Serial Number: 9QJ27KYM looks fine. (disk6 /dev/sdf ? ) Serial Number: MJ1311YNG5SD3A Firmware Version: MEAOA800 looks good. ( disk7 /dev/sdi ) Device Model: Hitachi HDS5C3030ALA630 Serial Number: MJ1311YNG5T15A looks good.
June 28, 201214 yr Author i am back again with the exact same issue. I precleared all drives and started from scratch 2 days ago as per your advice Joe and I've run into the read-only file system errors again. Is this now likely due to a faulty drive? It surely can't be something I'm doing? syslog.zip
June 28, 201214 yr Post a SMART report for the cache drive (currently sda). How are you shutting down or restarting? Have you run a memtest (preferably overnight)?
Archived
This topic is now archived and is closed to further replies.