August 18, 201015 yr We had some computer changes here, so my server's been offline (and partially dismantled) the last few days. I put it back together and rebooted yesterday (no hardware changes), and both disks read and wrote fine (~50Gb to each last night). But since this morning, it's refusing to write to \\media\disk2\tv\myShow but \\media\disk3\tv\myOtherShow is fine. Likewise \\media\tv\myShow won't write, but \\media\tv\myOtherShow is fine. There's plenty of room on both disks (~250GB free on each), they're both set to high water, and split level=1. Both read with no problems, and a parity check overnight was fine. A shutdown and restart hasn't made any difference. If I try to write to it using XP I get: "Access is denied. Make sure the disk is not full or write-protected and that the file is not currently in use." Windows 7 tells me: "You need permission to perform this action." Can anybody help me sort this out?
August 18, 201015 yr It's as I suspected, you have a filesystem corruption issue. Aug 18 11:25:33 Media kernel: REISERFS error (device md2): reiserfs-2025 reiserfs_cache_bitmap_metadata: bitmap block 33226752 is corrupted: first bit must be 1 Aug 18 11:25:33 Media kernel: REISERFS (device md2): Remounting filesystem read-only Aug 18 11:25:33 Media kernel: REISERFS warning (device md2): clm-6006 reiserfs_dirty_inode: writing inode 10730 on readonly FS You will need to go through the following steps: http://lime-technology.com/wiki/index.php?title=Check_Disk_Filesystems cd samba stop umount /dev/md2 reiserfsck --check /dev/md2 The above steps with brief explaination... cd [this will make sure you are in the /root directory] samba stop [all your shares will disappear from network] umount /dev/md2 ['md1' corresponds to disk1, 'md2' to disk2, etc. note it is 'umount', not 'unmount'] reiserfsck --check /dev/md2 [answer with the word Yes when prompted, do not type yes or YES, but Yes (capital Y and lower case es)] At the conclusion of the reiserfsck --check command, a report will be output. If errors are detected, this report may specify an additional action to take.
August 18, 201015 yr Author OK, it's done - it came back with: 1 found corruptions can be fixed when running with --fix-fixable Should I run that as per the other post you linked to? I've obviously lost some data - how can I tell what/how much?
August 18, 201015 yr Yes. Do what it tells you to; in this case you say it's: reiserfsck --fix-fixable /dev/md2 You may not have lost any data at all. It depends on what it can fix.
August 18, 201015 yr Author Bingo. It fixed it all, and says there's no lost files. Thanks a lot for your help, BRiT!
September 8, 201015 yr Hi, sorry for hijacking this thread but I'm having a similar problem. I also JUST upgraded to version 4.5.6 last night from 4.3 last night. I was unable to write to disk2, citing permission issues. I went to the console and saw a number of REISERFS errors on the screen. I ran reiserfsck as detailed in the link provided, and the result was: 8 found corruptions can be fixed only when running with --rebuild-tree I wanted to check here before proceeding. I captured a syslog, but it is too large to attach. I also didn't capture the log after running reiserfsck...is that necessary to troubleshoot?
September 8, 201015 yr Hi, sorry for hijacking this thread but I'm having a similar problem. I also JUST upgraded to version 4.5.6 last night from 4.3 last night. I was unable to write to disk2, citing permission issues. I went to the console and saw a number of REISERFS errors on the screen. I ran reiserfsck as detailed in the link provided, and the result was: 8 found corruptions can be fixed only when running with --rebuild-tree I wanted to check here before proceeding. I captured a syslog, but it is too large to attach. I also didn't capture the log after running reiserfsck...is that necessary to troubleshoot? You need to follow the instructions as described in the wiki for running reiserfsck on that drive. http://lime-technology.com/wiki/index.php?title=Check_Disk_Filesystems Typically, if the OS detects a corrupt file-system it will mount it as read-only to prevent further corruption. That is probably why you got "permission denied"
September 8, 201015 yr Thanks for the response, Joe. I ran reiserfsck, which is what gave me the "8 found corruptions can be fixed only when running with --rebuild-tree" response. I saw in the Check_Disk_Filesystems link that it said: Important Note!!! Do NOT run reiserfsck with the --rebuild-sb or --rebuild-tree switches, unless you are instructed to, by the instruction of a previous run of reiserfsck, or by an expert user! They are last-resort options, to repair a severely damaged Reiser file system, and recover as much as possible. They almost always leave the file system in worse shape than it originally was! At this point should I just go ahead and do the reiserfsck --rebuild-tree or should I take any precautions first?
September 8, 201015 yr Thanks for the response, Joe. I ran reiserfsck, which is what gave me the "8 found corruptions can be fixed only when running with --rebuild-tree" response. I saw in the Check_Disk_Filesystems link that it said: Important Note!!! Do NOT run reiserfsck with the --rebuild-sb or --rebuild-tree switches, unless you are instructed to, by the instruction of a previous run of reiserfsck, or by an expert user! They are last-resort options, to repair a severely damaged Reiser file system, and recover as much as possible. They almost always leave the file system in worse shape than it originally was! At this point should I just go ahead and do the reiserfsck --rebuild-tree or should I take any precautions first? You have been instructed by reiserfsck.
September 9, 201015 yr Just finished the --rebuild-tree. It moved about 8000 files to lost&found, and another 3000 were unrecoverable. The disk that I ran reiserfsck on (disk2) was recently replaced because I thought the previous disk was dead. Long story short, it looks like I may have been too hasty replacing the disk, and the previous disk may still be operational with all of the data on it. Is there any drawback to putting the old disk in and clicking "restore"? If the disk is working, it should have the exact same data that was on the current disk2 before running reiserfs and I could rebuild parity from it. If it is dead, I assume I could stop rebuilding parity, put back in the reiserfs'd disk, and restore again?
September 9, 201015 yr I hooked up the old drive so I could get a SMART report on it and see if it's even worth attempting to merge it back into the system. One thing I should note is that the old drive is smaller than the one with the bad file tree. Just from having it connected to the system for the short time, I noticed a potential problem. The system thought I was trying to upgrade a disk and flagged it as being too small. There weren't any options to restore even if I wanted to. smartOLDdisk2.txt
September 9, 201015 yr I hooked up the old drive so I could get a SMART report on it and see if it's even worth attempting to merge it back into the system. One thing I should note is that the old drive is smaller than the one with the bad file tree. Just from having it connected to the system for the short time, I noticed a potential problem. The system thought I was trying to upgrade a disk and flagged it as being too small. There weren't any options to restore even if I wanted to. The button labeled as "restore" was removes several versions ago. It was replaced with a command line command initconfig To use it, "Stop" the array log in via telnet or on the system console type: initconfig (Respond to its prompt) Back on the web-browser... refresh it and all the indicators should turn blue.
Archived
This topic is now archived and is closed to further replies.