hi,
i have a tricky situation here and and don't know where to start with
the solution.
i happened to set up a linux system, but i had to install the base
system on another machine. i used e2fsprogs 1.30-WIP (30-Sep-2002) for
this.
now the machine is up and running, but i have to use e2fsprogs 1.27
(8-Mar-2002). it's a debian system and i wanted to stay sane & stable,
hence using an older version of e2fsprogs.
but now i tried to check my fs and all i get is:
root@sheep:/# fsck.ext3 -v /dev/sda8
e2fsck 1.27 (8-Mar-2002)
fsck.ext3: Filesystem has unsupported feature(s) (/dev/sda8)
e2fsck: Get a newer version of e2fsck!
root@sheep:/#
##uuh, 1.30-WIP must be using some features to create a fs, which 1.27
##does not understand.
##so i grabbed 1.30-WIP and compiled the tools by myself (i can't use
##the binary-compiled version of 1.30-WIP because i would have to
##upgrade my libc6 -- and i don't wanna do this)
##
##now, weird things happen:
root@sheep:/# /usr/local/sbin/fsck.ext3 -v /dev/sda8
e2fsck 1.30-WIP (30-Sep-2002)
/dev/sda8: clean, 615/48960 files, 31719/195568 blocks
### ok, but didn't trust this 'clean':
root@sheep:/# /usr/local/sbin/fsck.ext3 -fv /dev/sda8
e2fsck 1.30-WIP (30-Sep-2002)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 3A: Optimizing directories
Optimizing directories: 16430
Pass 4: Checking reference counts
Pass 5: Checking group summary information
/dev/sda8: ***** FILE SYSTEM WAS MODIFIED *****
615 inodes used (1%)
17 non-contiguous inodes (2.8%)
# of inodes with ind/dind/tind blocks: 251/10/0
31720 blocks used (16%)
0 bad blocks
0 large files
519 regular files
87 directories
0 character device files
0 block device files
0 fifos
6 links
0 symbolic links (0 fast symbolic links)
0 sockets
--------
612 files
### fs was modified? what's going on here?
### a 3rd run gives:
root@sheep:/# /usr/local/sbin/fsck.ext3 -fv /dev/sda8
e2fsck 1.30-WIP (30-Sep-2002)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Problem in HTREE directory inode 16430: node (0) has bad limit (31744)
Clear HTree index<y>?
### well, and these HTREE issues are continuing (sp?) to infinity.
that means, in *every* fs-check there are always HTREE issues, and i
don't know if i should fix these by typing "y" or so.
what are your advises on this?
to re-check the issue, i created an 80MB zero file, created a ext3fs
with 1.30-WIP, tried to check with 1.27 and as assumed i got an error
about "e2fsck too old". checking and re-checking this file with
1.30-WIP
gives no errors, when no files are in it.
but loop-mounting this file and just copying /var/lib to this loop-dir,
unmounting it and checking it again with 1.30-WIP gives HTREE errors again.
please give me a hint on what to do.
in daily usage i got no errors and no anormalities in using the fs but
and don't wanna wake up one day and see everything screwed up.
Thanks in advance,
Christian.
--
############ Christian ##############
######## evilninja@gmx.net ##########
#####################################
what i forgot to say: i am using kernel 2.4.18 (debian version, called 2.4.18-5, only a few patches went in from these people), i applied the freeswan patch (IPSec) with no errors. gibc is 2.2.5, arch is ppc/PReP. these machines have ECC RAM, so i don't *believe* it's a RAM issue here. scsi-controller is a sym53c825a (yes, driver is Version 1, not Version 2) even more information available, of course. Thanks, Christian. -- ############ Christian ############## ######## evilninja@gmx.net ########## #####################################
On Jan 23, 2003 14:11 +0100, Christian wrote:> now the machine is up and running, but i have to use e2fsprogs 1.27 > (8-Mar-2002). it's a debian system and i wanted to stay sane & stable, > hence using an older version of e2fsprogs. > > but now i tried to check my fs and all i get is: > > root@sheep:/# fsck.ext3 -v /dev/sda8 > e2fsck 1.27 (8-Mar-2002) > fsck.ext3: Filesystem has unsupported feature(s) (/dev/sda8) > e2fsck: Get a newer version of e2fsck! > root@sheep:/# > > ##uuh, 1.30-WIP must be using some features to create a fs, which 1.27 > ##does not understand. > ##so i grabbed 1.30-WIP and compiled the tools by myself (i can't use > ##the binary-compiled version of 1.30-WIP because i would have to > ##upgrade my libc6 -- and i don't wanna do this) > ## > ##now, weird things happen: > > root@sheep:/# /usr/local/sbin/fsck.ext3 -v /dev/sda8 > e2fsck 1.30-WIP (30-Sep-2002) > /dev/sda8: clean, 615/48960 files, 31719/195568 blocks > > ### ok, but didn't trust this 'clean': > > root@sheep:/# /usr/local/sbin/fsck.ext3 -fv /dev/sda8 > Pass 1: Checking inodes, blocks, and sizes > Pass 2: Checking directory structure > Pass 3: Checking directory connectivity > Pass 3A: Optimizing directories > Optimizing directories: 16430 > Pass 4: Checking reference counts > Pass 5: Checking group summary information > > /dev/sda8: ***** FILE SYSTEM WAS MODIFIED ***** > > ### fs was modified? what's going on here? > ### a 3rd run gives: > > root@sheep:/# /usr/local/sbin/fsck.ext3 -fv /dev/sda8 > e2fsck 1.30-WIP (30-Sep-2002) > Pass 1: Checking inodes, blocks, and sizes > Pass 2: Checking directory structure > Problem in HTREE directory inode 16430: node (0) has bad limit (31744) > Clear HTree index<y>?Get a newer version of e2fsck (I think 1.32 is available). The 1.30-WIP version had a bug in it when regenerating htree indexes. Also, I'd suggest disabling htree entirely for now, since (a) your kernel doesn't support it yet, and (b) there are still some bugs in kernels that _do_ support it. Sadly, htree is enabled by default with newer e2fsprogs (even 1.32 I think) so it is kind of a pain. Cheers, Andreas -- Andreas Dilger http://sourceforge.net/projects/ext2resize/ http://www-mddsp.enel.ucalgary.ca/People/adilger/
### writing this mail again, because there were mail-adress issues.
### my fault, sorry.
hi,
i have a tricky situation here and and don't know where to start with
the solution.
i happened to set up a linux system, but i had to install the base
system on another machine. i used e2fsprogs 1.30-WIP (30-Sep-2002) for
this.
now the machine is up and running, but i have to use e2fsprogs 1.27
(8-Mar-2002). it's a debian system and i wanted to stay sane & stable,
hence using an older version of e2fsprogs.
but now i tried to check my fs and all i get is:
root@sheep:/# fsck.ext3 -v /dev/sda8
e2fsck 1.27 (8-Mar-2002)
fsck.ext3: Filesystem has unsupported feature(s) (/dev/sda8)
e2fsck: Get a newer version of e2fsck!
root@sheep:/#
##uuh, 1.30-WIP must be using some features to create a fs, which 1.27
##does not understand.
##so i grabbed 1.30-WIP and compiled the tools by myself (i can't use
##the binary-compiled version of 1.30-WIP because i would have to
##upgrade my libc6 -- and i don't wanna do this)
##
##now, weird things happen:
root@sheep:/# /usr/local/sbin/fsck.ext3 -v /dev/sda8
e2fsck 1.30-WIP (30-Sep-2002)
/dev/sda8: clean, 615/48960 files, 31719/195568 blocks
### ok, but didn't trust this 'clean':
root@sheep:/# /usr/local/sbin/fsck.ext3 -fv /dev/sda8
e2fsck 1.30-WIP (30-Sep-2002)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 3A: Optimizing directories
Optimizing directories: 16430
Pass 4: Checking reference counts
Pass 5: Checking group summary information
/dev/sda8: ***** FILE SYSTEM WAS MODIFIED *****
615 inodes used (1%)
17 non-contiguous inodes (2.8%)
# of inodes with ind/dind/tind blocks: 251/10/0
31720 blocks used (16%)
0 bad blocks
0 large files
519 regular files
87 directories
0 character device files
0 block device files
0 fifos
6 links
0 symbolic links (0 fast symbolic links)
0 sockets
--------
612 files
### fs was modified? what's going on here?
### a 3rd run gives:
root@sheep:/# /usr/local/sbin/fsck.ext3 -fv /dev/sda8
e2fsck 1.30-WIP (30-Sep-2002)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Problem in HTREE directory inode 16430: node (0) has bad limit (31744)
Clear HTree index<y>?
### well, and these HTREE issues are continuing (sp?) to infinity.
that means, in *every* fs-check there are always HTREE issues, and i
don't know if i should fix these by typing "y" or so.
what are your advises on this?
to re-check the issue, i created an 80MB zero file, created a ext3fs
with 1.30-WIP, tried to check with 1.27 and as assumed i got an error
about "e2fsck too old". checking and re-checking this file with
1.30-WIP
gives no errors, when no files are in it.
but loop-mounting this file and just copying /var/lib to this loop-dir,
unmounting it and checking it again with 1.30-WIP gives HTREE errors again.
please give me a hint on what to do.
in daily usage i got no errors and no anormalities in using the fs but
and don't wanna wake up one day and see everything screwed up.
Thanks in advance,
Christian.
#########
what i forgot to say:
i am using kernel 2.4.18 (debian version, called 2.4.18-5, only a few
patches went in from these people), i applied the freeswan patch (IPSec)
with no errors.
gibc is 2.2.5, arch is ppc/PReP. these machines have ECC RAM, so i don't
*believe* it's a RAM issue here. scsi-controller is a sym53c825a (yes,
driver is Version 1, not Version 2)
even more information available, of course.
--
############ Christian ##############
######## evilninja@gmx.net ##########
#####################################