Time to migrate ext2 to ext3. We had many dramas with ext2, it's O.K. for a
single-user machine perhaps, but
you still can get data loss and f/s probs. Particularly as our machines were
still waiting on UPS's and getting powered
off frequently! We now get machines having power dropped regularly and they
don't miss a beat.
We're using EXT3 (Linux journalling FS) and software RAID-1 at 120 sites
with great success. ReiserFS is
apparently good for large amounts of small files. XFS is another option
which we've considered and handles large
files well. Linux has been such a success we will continue rolling it out in
future. We're using Redhat 7.1/7.2 for
our setup and it's been a dream to look after.
And you've got to love the upgrade path for EXT3. Recompile kernel with
EXT3, reboot. Install
a couple of updated RPMs.
(Redhat 7.1 updates required, rsync is just in there because I wanted it,
used for remote secure ssh2 file transfers):
RPM_LIST="e2fsprogs-1.25-2.i386.rpm mount-2.11n-4.i386.rpm
rsync-2.5.1-2.i386.rpm util-linux-2.11f-17.i386.rpm"
Process is:
Recompile kernel to support EXT3. (Ours supports EXT2/EXT3 and XFS for
possible future migration)
Install kernel, install kernel modules, change lilo config and re-run lilo
to update.
Reboot to new kernel.
tune2fs -j /dev/hdxx (convert EXT2 to EXT3 by creating journalling file)
tune2fs -i 0 -c 0 (disable periodic/count checking as the journalling will
never use it, and it's dangerous now)
Change /etc/fstab to read ext3 instead of ext2.
Reboot to your all new journalling Linux system :)
This was run on 120 _live_ ext2 systems successfully as follows:
# Set each filesystem to journal. Disable the periodic checking as this is
# not required with EXT3 journalling.
df -k | egrep -vy "filesystem|cdrom" | awk {'print $1'} |
while read -r INVAL
do
echo "Implementing journalling for filesystem: ${INVAL}"
/sbin/tune2fs -j ${INVAL}
echo "Disabling filesystem checking for filesystem: ${INVAL}"
/sbin/tune2fs -i 0 -c 0 ${INVAL}
done
cp /etc/fstab /etc/fstab.orig
cp /tmp/fstab /etc/fstab
# Reboot
> -----Original Message-----
> From: Walker, Thomas [SMTP:Thomas.Walker AT EMICAP DOT COM]
> Sent: Monday, March 25, 2002 11:47 PM
> To: ADSM-L AT VM.MARIST DOT EDU
> Subject: Re: linux TSM4.2.1: backup problem w/ ext2
>
> You won't usually get a reliable result from e2fsck without unmounting the
> filesystem first. Since's it appears that this is the /home partition,
> maybe
> you should logout all normal users and umount /home and rerun e2fsck to
> make
> sure there REALLY are no errors on the filesystem. ReiserFS works
> beautifully on 4.2.1 btw and there is no need to run a fs check on
> journaling file systems (99.9999% of the time :-) )
>
> -
> Tom Walker
>
>
> -----Original Message-----
> From: Christian Glaser [mailto:christian.glaser AT DLR DOT DE]
> Sent: Friday, March 22, 2002 9:51 AM
> To: ADSM-L AT VM.MARIST DOT EDU
> Subject: linux TSM4.2.1: backup problem w/ ext2
>
>
> hello all,
>
> i have a problem with my linux-client, TSM-server is 4.1.4
> --------------------
> [urmel@moby urmel]$ uname -a
> Linux moby.ae.go.dlr.de 2.4.9-6custom #1 SMP Tue Oct 30 11:41:02 CET
> 2001 i686 unknown
>
> [urmel@moby urmel]$ dsmc
> Tivoli Storage Manager
> Command Line Backup Client Interface - Version 4, Release 2, Level 1.0
> (C) Copyright IBM Corporation, 1990, 2001, All Rights Reserved.
> ---------------------
>
> i recently noticed this entries in my dsmerror.log:
> ------------------
> 03/21/2002 16:51:05 TransErrno: Unexpected error from lstat, errno = 9
> -----------------
> after some investigation it was clear that some files/dirs never got
> backed up following this error-msgs.
> this error seems to inidcate a bad filedescriptor.
>
> so i made a filesystemcheck assuming the error there.
> ----------------------
> [root@moby /root]# fsck /dev/hda5
> Parallelizing fsck version 1.23 (15-Aug-2001)
> e2fsck 1.23, 15-Aug-2001 for EXT2 FS 0.5b, 95/08/09
> /dev/hda5 is mounted.
>
> WARNING!!! Running e2fsck on a mounted filesystem may cause
> SEVERE filesystem damage.
>
> Do you really want to continue (y/n)? yes
>
> /home was not cleanly unmounted, check forced.
> Pass 1: Checking inodes, blocks, and sizes
>
> Pass 2: Checking directory structure
> Pass 3: Checking directory connectivity
> Pass 4: Checking reference counts
> Pass 5: Checking group summary information
> /home: 33232/2424832 files (43.4% non-contiguous), 4580356/4843589
> blocks
> --------------------------
> nothing suspicious here.
>
> i thought the error must be on client-system side as it is an error by the
> 'lstat' system function. but how come that the fsck went fine and no other
> programs are compaining - at least as far as i noticed it yet.
>
> any hel appreciated.
>
>
> best regards /
> mit freundlichen gruessen
> christian glaser
> _____________________________________________________________________
> T-systems Solutions for Research GmbH
> c/o DLR - 8234 oberpfaffenhofen tel: ++49 +8153/28-1156
> e-mail: christian.glaser AT debis-sfr DOT de fax: 28-1136
>
> This e-mail including any attachments is confidential and may be legally
> privileged. If you have received it in error please advise the sender
> immediately by return email and then delete it from your system. The
> unauthorized use, distribution, copying or alteration of this email is
> strictly forbidden.
>
> This email is from a unit or subsidiary of EMI Recorded Music, North
> America
**************************************************************************
Bunnings Legal Disclaimer:
1) This document is confidential and may contain legally privileged
information. If you are not the intended recipient you must not
read, copy, distribute or act in reliance on it.
If you have received this document in error, please telephone
us immediately on (08) 9365-1555.
2) All e-mails sent to and sent from Bunnings Building Supplies are
scanned for content. Any material deemed to contain inappropriate
subject matter will be reported to the e-mail administrator of
all parties concerned.
**************************************************************************
|
