Ubuntu Installation :: Ext3 Vs Ext4 When Upgrading To 10.04 From 8.04?
Sep 7, 2010
I am going to do network upgrade of my 8.04 to 10.04 LTS. I have a dual boot with Vista, and currently I can read Ubuntu partition from Vista using Ext2fsd. However, I read that using Ext2fsd with ext4 (default in 10.04) is problematic because of "extents".
My questions: Will the upgrade convert my current ext3 partition into an ext4 one? How can I keep the ext3 partition? The release notes for 10.04 say:
The simplest way to select a different file system such as ext3 at installation time is to add the partman/default_filesystem=ext3 boot parameter when starting the installer.
Is that possible, I mean when I upgrade F10 to F11 with yum upgrade is there a way to 'upgrade' the filesystem to ext4 for example (with the exception of boot partition)? Or I have to reinstall fedora like new?
While changing the filesystem can I do it by parts? what I mean is for example: I have 2 partitions like '/' and '/home' with ext3, so I backup data in '/home', change '/' to ext4 then mv files from '/home' to '/' and change '/home' to ext4 and finally mv those files from '/' to '/home'. Is that possible?
I had 5.4 machine. Upgraded to 5.5 today via yum upgrade. All went fine. Rebooted. Wanted to convert root partition to ext4 (I have three partitions: /boot, / and swap). All of them on software RAID 1 (root is /dev/md2). I did the following for converting
yum install e4fsprogs tune2fs -O extents,uninit_bg,dir_index /dev/md2 nano /etc/fstab # I indicated here that my /dev/md2 is of ext4
The laptop I regularly use is out of use after having tried to upgrade from ext3 to ext4, following the instructions that can be found at namely, Step 2: Upgrading to Ext4Well, the first step was done without problems, but when I reached "Mount your filesystem", the problems started:When typingsudo mount -t ext4 /dev/XXXX /mntwhat appears is:UNEXPECTED INCONSISTENCY; RUN fsck MANUALLYStupid me, I did nothin but jump to the next step:gksu gedit /mnt/etc/fstabWhere I only got a fully white screen with no written text at all, instead of the menu that should have appeared, as can be seen on that webpage.
However, after running Gparted again, it showed the extension as EXT4, not EXT3, so I thought the system had automatically upgraded. Then I proceeded to boot the system, and this is where the massive failure appearedAfter seeing the Ubuntu logo for a while, the system directs me to a blank screen with the message:Mount of filesystem failed.dc5123-ec01-4438-8bd6-cb85bb080f87A maintenance shell will now be started.CONTROL-D will terminate this shell and re-try.root@hihihi-laptop:#So I type CTRL-D:
mountall start/startingfsck from until-linux-ng 2.16 /dev/sda1: One or more block group descriptor checksums are invalid. FIXED /dev/sda1: Group descriptor 0 checksum is invalid
I have been able to get most of the way through the process of changing from using ext4 back to using ext3, but something is not quite right so my system does not boot properly.
I have a system that was running Karmic Koala 9.10 as a server (no graphical environment). I had two drives using RAID1 with LVM on top, where the logical volumes of oldvg (old volume group) were using mostly ext4. /boot was not part of the RAID: it's on a separate physical drive and uses ext2.
I recently added two more drives and used RAID1 and LVM, and made all lv partitions (/, /usr, /var, /tmp, /opt, /home, /srv) ext3. I used rsync to duplicate the contents onto the logical volumes of newvg (new volume group). I was careful with rsync's option switches, and this part seems to be fine.
I also edited (the new) /etc/fstab and changed the UUIDs of the seven mount points to point to the logical volumes that are part of newvg instead of oldvg, and added new entries to (the new) /boot/grub/menu.lst to refer to newvg in addition to those that I left around to refer to oldvg.
This wasn't sufficient: rebooting here failed, but I went in with a rescue disk, and first updated /boot/grub/device.map to include the new physical drives. I then mounted all the new logical volumes, mounted boot also at its proper place, and entered a chroot of the new system as it should be mounted. Once there, (and after making a backup of /boot) I ran "update-initramfs -k all -c" to rebuild the initrd images that were stored on /boot. Finally, I also edited /etc/mtab so that the two entries that referred to oldvg now refer to newvg instead.
Now, the machine begins to boot from newvg, but the console text includes messages like:
And a bit later,
Now, at this shell if I type mount, I see:
I am actually confused as to why there are only entries for /root and /var in /etc/mtab, actually, instead of entries for all of the main mount points. I am thinking it must be part of the boot staging process, because there are entries for newvg-usr, newvg-tmp, etc. in /etc/fstab.
When I type any of pvdisplay, vgdisplay, or lvdisplay, I get
In fact, even if I run lvm, I get a similar error:
However, if I go back to the rescue cd, pvdisplay, vgdisplay, and lvdisplay do show that all of the partitions from both the old and new volume groups are available.
I have Debian Squeeze with linux-2.6.32-5-amd64Ok, so what I did is first changed fstab ext3 to ext4, rebooted, everything looked good. Then I enabled extents and other ext4 features on my root and home partitions using commands
tune2fs -O extents,uninit_bg,dir_index /dev/sda3 tune2fs -O extents,uninit_bg,dir_index /dev/sda4 then I ran e2fsck -fDC0 /dev/sda3
It has crashed or failed to install on a known good system for the 5th time now. I have never had that problem with fedora before. I am now installing with ext3 instead of ext4. I am hoping for better success with it and will update. I was able to try MinGW with WINE before the first crash and I liked what I saw there. Only hoping to be able to get a good install this time.
Then I convert the / partition filesystem by the following steps:login as root user in multi user mode read [url], and execute tune2fs -O extents,uninit_bg,dir_index /dev/mapper/VolGroup01-LogVol00 modify /etc/fstab, change the type of / to ext4
reboot (because fsck say running it on a mounted filesystem can cause filesystem damage, so i decide to reboot to single user mode first. maybe it's a mistake here) try to boot F11 to single user mode, failed
reboot from a SystemRescueCD-1.2.0 LiveCD run e2fsck -fpDC0 /dev/mapper/VolGroup01-LogVol00 in SystemRescueCD-1.2.0 LiveCD, no error reported
reboot try to boot F11 normally(multi user mode), but it failed at: EXT3-fs: dm-0: couldn't mount because of unsupported optional features (40).
I tried rescue mode of Fedora-11-i386-netinst.iso after these steps above. The / partition can be found and mounted to /mnt/sysimage correctly. I can read/write files in / partition, and i can even yum new softwares in rescue mode, but it just can't be mounted when booting.
I run an upgrade and an update on a lucid lynx beta 2. --- got no problems. but about the filesystems i have some questions because it seems for me that at every system boot the system will run an fsck. somtimes it's shown up, somtimes not. but in /var/log/messages and in syslog
I have always following messages ( occured in beta 2 too ).
But first before i continue - here my disk layout:
And here my filesystem types:
This is my problem because those values are seems to be static ! ( note: this partiton is mounted but not in use ) and last not least: the drive is an external usb scsi disk. but on the other side lucid lynx is running fine on my box.
I have a drive with ext3 that was my old /home drive.When I moved to 9.04 I went with ext4 and just linked to my old home drive.My old ext3 drive is becoming unstable so I want to back it up.I have a 500gb ata drive (unfortunately not a sata drive) thatto wipe and put 10.04 on. I want to move my data from this backup driveIm going to back up to, to the new 10.04 until I get another hd as abackup drive.My question is, I know there is some issue between ext3 and 4 and I dontwant to get into a bind. What would be the best format for moving
Is there guide for converting ext3 to ext4 on Fedora? I use Fedora 12 which is regularly updated. How safe is procedure for data, I have only one ext3 partition on disk which has one ntfs and that ext3 partition (and also one small swap partition).
how to convert ext3 to ext4? I'd like to convert partitions which I use for virtual machines (vmware-server and virtualbox). I use Ubuntu 9.10 as vmware-server host and Gentoo as virtualbox host.
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've recently installed Ubuntu 10.10 on a machine, unfortunately, the hard drive was ext3 partitioned. Is there a way of converting this partition to ext4 without having to re-format and hence reinstall the entire OS ?
I need to convert my server(10.04) to ext4 from ext3 and want to make sure I am going to do this the correct way.
Edit fstab to ext4
Code: sed -i s/ext3/ext4/ /etc/fstab Run the tune2fs command, the filesystem is mounted is this ok? Code: tune2fs -O extents,uninit_bg,dir_index /dev/md0
The ext4 features are well listed in the Ext4 Howto.
Most of these features seem to be aimed at newer systems, having petabytes disk drives and handling huge files -- neither of which applies to my system.
Can a modest 3 year old system, having only two 500GB HDDs, used mostly for Android development using Eclipse (64-bit), benefit from choosing ext4 over ext3?
I've heard that ext4 has better performance, but that it will also eat my hard drive. Has this problem been fixed? What would be a safer bet on a squeeze box? Is there even anything to worry about?
I'm wondering if anyone knows what will be Squeeze's default filesystem. Will it be the proven ext3 or the newer (sometimes faster, sometimes slower) ext4?
I currently have ext4 and I have nothing to complain about. In fact, my overall experience has been very positive. Ext4 is definitely faster when fsck runs during boot.
What would be the cons of having ext4 as default in Squeeze?
I have installed ubuntu to my pc. i made 3 partitions. one for system, one for data and one for swap. two of them were ext4. after some time i have reinstalled ubuntu again. but this time i didn't put to format the second partition, but just mount it using ext4. after that i cannot open my files. checked with gparted shows that 2GB used, but with df 188MB. and in properties writes ext3/ext4 filesystem. i used chown, chgrp but didn't help. please help, these data are ver important. i cannot lose them.
How can I format a USB hard drive to ext3/ext4 or whatever file format and have full permission to read, write and execute all files afterwards? When using the command line (as ROOT of course) mkfs.ext3 /dev/sdb? Restricts the rights to ROOT as does the procedure gParted. The man mkfs did not help much. Configuring the fstab- file is a bit of a hassle, so it would be nice, if there was an option to set the permissions "correctly" right from the beginning. Setting Ubuntu (I'm using Ubuntu 9.10) up, so that it mounts USB devices not as ROOT as default but giving all users all permissions seems to be really complicated, as a guy from my local LUG told me.
I compiled a kernel using a previous kernel config after I switched to the ext4 file system. Previously it was using ext3 and the kernel compiled and ran fine. I added support for ext4 to the config but when I went to boot I had a kernel panic. The error was "kernel can't mount vfs on (8,5)". Root is on sda5, I don't know what the 8 is. I started over, using mrproper and made a new config, but got the same error. I created an initrd with the ext4 file system but then the kernel said it couldn't mount root on ext3 because of unique options. (something along those lines) I booted back into the default kernel and saw that it had a similar error right after the bios check, but it loaded fine. I'm wondering why the kernel is saying that the ext4 file system is ext3.
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
Which file system uses Centos 5.6 by default, Ext3 or Ext4? I have installed on Ext3, it's upgrade from 5.5, but howto convert into Ext4 without damage or lost data?
I work with lots of big files >500MB. As ext3 does not provide the extents feature of ext4, I am concerned that some fragmentation may periodically affect disk performance. Noted I understand that using ext4 under these circumstances would be pointless, because the extents feature would be required to be disabled to allow compatibility using ext2fsd at this time, thereby removing any tangible ext4 benefit for my purposes.
Question Would an effective method of disk optimisation and anti-fragmentation be to simply copy all files from the ext3 media partition across to another entirely separate empty backup ext3 drive, before copying them all back over to the ext3 media partition again? Would this mean that when all the files are copied back again, they would be placed back in the 'correct and optimum order', so to speak? PS: I cringe in advance at the n00bs who are likely to reply to this thread with 'ext3 cannot become fragmented'!
I've been using full disk encryption with luks on two Dell laptops for about 2.5 years. One is 64 bit Fedora (upgraded to 12), another is 32 bit Ubuntu (upgraded to 9.10), both were ext3, until recently.Over time, performance degraded substantially, especially on Fedora, which was put to a heavier use. That laptop has 4 Gb RAM, two 2.5 GHz T9300 CPUs, and 56 Gb of free space.
It was especially unbearable after a reboot. Programs like firefox and thunderbird would take close to a minute to start when ran for the first time after a boot. The login process was painfully slow, and some Gnome applets (e.g. Tomboy notes, keyboard layout switcher) would fail to load on the first login, with an error. I experienced this problem on both laptops with full encryption. I had to log out and relogin to make the applets appear. I tried various boot and mount options and was thinking about switching to ecryptfs (encrypted home).
I also use 3 desktops with no encryption and a netbook with ecryptfs on /home, which all work fine. All are Dell, 2 Ubuntu and 2 Fedora. The Gnome applets problem seems to be due to slowness of the installs with the full disk encryption. The last thing I tried is to migrate ext3 to ext4. I also converted /home, /usr, /opt to extents, following[URL].. That seemed to do the trick. Gnome applets now load fine on both laptops, and startup time is back to tolerable.
Is this a typical experience: ext3 performance degradation with time and a much better performance with full disk encryption once ext3 is migrated to ext4?
