CentOS 5 Hardware :: Fsck Fails On Ext3 Filesystems?
Sep 15, 2010
I am attempting to run a fsck on a number of large ext3 filesystems. I am doing this proactively because I want to minimize reboot time and the filesystems are past the interval time of 6 months. When I run the command " fsck -f -y device" I get the following error on all of the filesystems-
fsck 1.39 (29-May-2006)
e2fsck 1.39 (29-May-2006)
fsck.ext3: Device or resource busy while trying to open /dev/mapper/mpath0p1
I am hosting a few customer servers now, all of which are virtual machines running on a CENTOS 5.x host. Each Centos host has a couple of extra drives. When I formatted them ext3 they automatically had a schedule of a full forced fsck after 6 months. Do I really need to do that check regularly? It results in a fairly large outage since my disks are each 1TB and there are up to three extra drives on each server. I try to reboot these servers every 6 months but this part adds a large amount of time to a routine reboot.
I have a serious problem in booting centos 5.4 x86 as shown in the attached picture.I tried to backup before using fsck command, but I could not make a backup of damaged lvm on hard drive.First I made a rescue centos at virtualbox, and installed centos 5.4 x86 on virtual hard disk.And I attatched damaged hard drive. S I can see this damaged hard drive's lvm as attached picture.Please let me know how to backup my files and to use "fsck.ext3 --rebuild-tree using livecd".
I Tarred and GZipped most of the data on one 1Tb partition and stuck the archives on a second 1Tb partition on a separate disk. I then proceeded to format the first partition with NTFS (from Linux.) The only problem is that I completely forgot that I had a CD drive and formatted sdc1 instead of sdd1! I began doing a full NTFS format and after a minute or two I cancelled it and decided to do a quick format. I then realized my mistake. I managed to find a copy of the superblock and began trying to recover the disk. fsck -t ext3 recognized the partition as NTFS but I luckily didn't have fsck.ntfs installed so it didn't touch it. I managed to get it working with fsck.ext3 (with -b,-B and -y) fsck.ext3 didn't mind that it was an NTFS partition.
Roughly how long will this take? It's running from Knoppix within a virtual machine to a USB hard drive which is 100% full. Days? Being that for a few minutes I attempted a full format am I going to end up with a bunch of corrupted archives? If I do end up with file corruption can anyone recommend a way of recovering the data / sorting it out? Is it likely to be just a few old files that are corrupt (It's my understanding that filesystems like to keep files in the same area on the disk to minimize the amount of head travel.) This might just be wishful thinking but as the filesystem fills up will ext3 put the newer files towards the end of the disk? If so then I'm hoping that a full NTFS format starts at the beginning of the disk.
I want the filesystem of my external drive to be checked periodically after a numer of mounts. I put 2 in the sixth colums of fstab for this partition
Code: /dev/sdb1/mnt/hdext3rw,dev,sync,user,noauto,exec,suid02 and I use the tune2fs to set the maximum mount count to 32. Code: tune2fs -c 32 /dev/sdb
now the mount count is 34 and the date of the last check is not recent, so apparently the auto fsck has not been performed. Probably because this partition is not mounted at start-up but I usually mount it manually.
My openSUSE 11.2 system has periodic running of fsck disabled for ext4 filesystems (Maximum mount count = -1, Check interval = 0). What is the reason for this? Is it because fsck is not necessary on a periodic basis with ext4, and only necessary when errors are detected? Or is it because fsck has problems working on ext4 filesystems?
The ext3 filesystems do have it set (Maximum mount count = 500, Check interval = 5184000 (2 months)). I would like to know why fsck is not set to be activated for ext4.
I have a centOS 5 box with 3ware 8 port raid cards.I run fsck.ext3 -y /dev/sdb1 and it shows as clean.But after writing to the FS for about 2 minutes, it becomes read-only.When I umont -l /data, and run the fsck I get that another program is using the system an I should wait.If I reboot the server , the array comes back as clean.
fsck.ext3: Unable to resolve 'UUID=theUUID' where "theUUID" (without the quotes) is the UUID
I believe this is caused by me trying to get lvm to use the external /boot because when I had unmounted the external /boot, it was creating a /boot in root. So, I booted a live cd and mounted the external /boot where /boot in the root volume is supposed to be. Basically, I think the problem is that I need to make my /boot (which is the only ext3 partition in the entire system and I want it that way) "relate itself" to the lvm root so that it boots into the system. As mentioned earlier, in the live CD, I made the external /boot mount itself in the root's /boot but I don't know how to tell the system to do this on its own while booting without my assistance. I chrooted from the live cd which involved a lot of tedious stuff but basically the important stuff I did were:
I have a problem partitioning an Hard Disk Drive on a server, and I hope someone can help me with this. Here is the system configuration: Operating System Linux localhost.localdomain 18.104.22.168-64.fc11.x86_64 #1 SMP Fri Sep 25 04:43:32 EDT 2009 x86_64 x86_64 x86_64 GNU/Linux
Hardware: RAID bus controller: Hewlett-Packard Company Smart Array G6 controllers (rev 01)
The system mounts one hard disk (120 Gb) with the OS, and four 1.5Tb Hard Disks mounted on RAID 10 for a total of ~3 Tb I need to create several partitions on this RAID drive, but I have some trouble doing it. I need a total of 10 partitions of different sizes:
My ubuntu stops when mounting system hdd. The screen display the following messages :
mountall:/etc/fstab: No such file or directory fsck from util-linux-ng 2.16 WARNING: couldn't open /etc/fstab: No such file or directory init: mountall main process (545) terminated with status 1 General error mounting filesystems. A maintenance shell will now be started. CONTROL-D will terminate this shell and re-try. udevd: can not read '/etc/udev/rules.d/z80_user.rules. Ubuntu: clean, 474879/24231936 files/28016581/96898047 blocks root@i7:~# exit_
I suspect the disk manager pysdm that i had just installed today and it had crash during the previous session. The /etc/fstab file does not exist anymore and i cant rename the fstab.bak because the disk is read-only even for my root user
I am travelling so would really appreciate a few tips that should hopefully get my pc running again. (I have internet access via a kiosk, and can burn a disk if necessary, but the smaller the download the better)I have a few ideas that with some help I should be able to fix the partition.1. Is there a simple way to use use grub, to mount all partitions read only and not to fsck them so I can logon in normal user and sudo and hopefully fix the parition?or 2. Download a small version of ubuntu or something that I can burn to CD (my machine cannot boot from usb) so that I can repair the machine.3. Download the ubuntu installer (alternative) and try and reinstall the necesary usr partition?
I've just migrated to Debian from Ubuntu on my server. So my first post here. I've had some problems and questions when installing and configuring Debian, now everything seems very smooth, but still I can't find answer to one question. My machine is headless and I allways use SSH to it. I have several partitions on machine so if one of them fails on booting time with fsck, I need to press CTRL+D combination to continue. In that case I need to go to the machine with monitor and keyboard and press that combination. Is there any way to automatically continue for example after 30 seconds ? Then I could log in to the system and check log files what is wrong. How can I do that ?
I have a 2TB file-system and when the machine reboots it fails the fsck, halts and goes into maintenance mode.Stats: I have have RHEL 5, 2.6.18 kernel, the file-system is an ext3. The file-system is on an EMC AX4 connected with fiber channel HBA.So far my reading tells me this should work because under 2.6 4TB is OK. Any ideas why this fails?If I take it out of the fstab file and mount it manually the boot is OK and the file-system behaves well. I can change the fsck check option in the fstab to 0 but I don't think I should have too. Everything I read says that 2TB ext3 file-systems are OK.
i am trying to get centos 5.4 installed and bootable on my 16gb flash drive, with persistent overlay using and ext3 formatted drive.i want to be able to boot into centos and be able to have all updates from yum, etc, saved when shutting down for my situation i cannot use vfat.
I had a server with CentOS 5.5 and Asterisk fail today most likely due to the power failure. I guess the stupid electrician turned the power ON and OFF multiple times while changing the burned out fuse without unplugging the PBX. It was set in CMOS to power back on when there is power outage so my two HDDs which were set in RAID-1 now shot.
Trying to start the system and it goes through some check and fixes and then tells me to do Ctrl+D or enter password and run fsck. I have tried both methods and I don't think they yeild anything.
Ctrl+D reboots and same store again with the system fixing and asking me to Ctrl+D again. Putting the password and doing "fsck -y" gives many errors and it fixes it and then it just keeps looping. I am afraid this is making it worst than better. So, after few loops I did few more restarts and just give up. Took out the HDDs and installed new ones and install CentOS+Asterisk so business runs as usual on Monday.
However, I am kind f*ked if I don't get the files out of these HDDs. This is my first time using fsck and also encountering this type of a problem. What is the normal procedure when fixing inodes misplaced in a case like this? and what should I do to at least be able to grab my files?
I have already connected one of the drives to Windows and used Disk Internals (which maps Linux drives) but I see only folders with no contents except for /tmp/ which has contents. While the HDDs were in the system I could actually browse files and they were fine. I am really keen to get the drives fixed but I am also afraid that bad usage of fsck might replace lots of files and burn it all.
I'm using CentOs 5.4 (2.6.18-164.15.1.el5 #1 SMP Wed Mar 17 11:30:06 EDT 2010 x86_64). I tested out ext4 on a partition for the last few months and it seems to work fine. The issue is that quotas dont seem to work correctly on it. Is there a way to revert back to ext3? Mainly the quota tools do not work on it.
I have a external usb 20Gb hdd. It was a ntfs partition but i could'nt mount it, so I try with a ext3 partition, because with centos it's mount automaticly, but... for some reason that I don't know, even with a ext3 partition my centos don't detect my external usb hdd, howerver I already try to mount it on ubuntu to see if it work, even on windows with some programm to recognize ext3, it was alright, but why can't i mount it on my centos?
With the release of CentOS 5.5 ext4 is considered stable in this distribution so I decided to migrate to it. Luckily I started from migrating fresh server with CentOS 5.5 using some instruction I found on the internet. I think I shouldn�t say, that I screwed the whole thing up ;) After about 6 hours cursing, kicking, and crying I solved the task and figured the correct sequence of actions. The small problem with migrating root partition is that you can�t unmount it BTW.
During migration task, I found, that CentOS 5.5 rescue mode is somewhat broken a little in terms of ext4 support. It can mount ext4 partitions successfully. But its e2fsprogs package (tune2fs, e2fsck etc.) doesnt see ext4 partitions and say, that superblock is corrupted on a partition once is converted to ext4 (at least it did it for me. May be I should force filesystem type with -t ext4 switch?). Keep in mind, that if you screw your system up too badly, you will not be able to run tune2fs and e2fsck on it from rescue modeBut you will still able to mount it if it is not corrupted badly. In all below examples,Boot your system normally and login as root. Upgrade kernel if you wish (I usually use yum upgrade to upgrade all on new machines). Then upgrade/install some other packages
I am unable to boot my server and it is erroring with the message "Kernel panic - not syncing: Attempted to kill init!". Here are the details:
I have a brand new Dell PowerEdge T110 server with Quad-Core Xeon CPU and 2GB 1333MHz RAM. I have installed a hard disk (SATA, 7.2rpm) on this machine that I removed from another Dell PowerEdge T100 server with Dual-Core Xeon CPU that was running software RAID 1.
Now when I boot the new T110 server, it displays the following error messages code...
I am running CentOs v5.5 x86-64. If I put the hard disk back into T100, it boots fine. I do not know why it is not working with T110 server.
I am going to install CentOS 5 on my pc , for3d modeling , and i have some doubts about partitioning the system.
I know that i can have /, /var , /tmp , /home and other directories as individual partitions. I know that i can use ext3 , or xfs ( i know that anaconda doesn't let to use xfs, but i know a way of migrate partitions to xfs after installing) or ext4 , and others , without problems.
But i could for example, once installed all with ext3 , have for example / on ext3 , and /var and /home on xfs.
For me it would be very interesting have /home in a another partition and as xfs (migrating it later from ext3) as mostly i will save in /home a lot of big files ,images,dvd isos,renders , binary maya files and like, and i see a advantage of use xfs with /home.
But a lot of people also use xfs for / as well , but as / has a lot of small files , perhaps use xfs on /var and some other directories, could be good (because xfs performance deleting files or other features), and leave / as ext3 for example.
So is it possible to have different filesystem type for some directories under / ,(for example /var and /home as xfs, the rest as ext3 in / ) , without having any problem ?
(not counting using xfs for a directory that won't see improved his performance because has a lot of static small files or like)
I have configured a "Syslog" server on /var directory as a separate ext3 partition - to receive the logs and events from the clients & the firewall as well. The directory needs to grow dynamically as the logs are populated. Is there a way i can make the filesystem grow dynamically as and when the directory is full.
I have more than a dozen servers running CentOS (mostly supermicro servers) everytime i do mkfs.ext3 or mkfs.ext4, it always hangs the server (nothing is responding, not even the terminal) can't even do Ctrl+C to kill it. but the funny thing is, if i leave it on for hours, it'll finish eventually.
this puzzles me, i've tried so many different SATA drives and different servers, different kernels, different CentOS. for example purposes, my CentOS 5.3 runs 2.6.18-92.1.13.el5 kernel. i also have 5.4 and 5.5. mkfs.ext3 works fine on SAS, but always hangs on SATA. here's where it usually hangs Writing inode tables: 2126/7453