headers
mt | 8 Nov 2003 03:32
Picon

Report: mounting JFS fails unless log got replayed


Hello all,

as reported earlier on this mailinglist I had problems
mounting an JFS partition - it failed to mount when
connected to an external device.

Meanwhile I managed to make the problem reproduce & fixable:
mounting the JFS partition fails when the partition's journal
was not already replayed by fsck.

here the related lines from /etc/fstab
#/dev/hdg1      /mnt/sda1       jfs     defaults                1 2
#/dev/hdg2      /mnt/sda2       jfs     defaults                1 2
/dev/sda1       /mnt/sda1       jfs     defaults,noauto         0 0
/dev/sda2       /mnt/sda2       jfs     defaults,noauto         0 0

Note: I'm using an external IDE-Firewire bridge, so it's an SCSI-device
to be mounted - anyway the IDE-HD is still the same.
It's obvious that the on-boot scanning and error-attempt values have to
be zero for an externally connected device that may be plugged in
at any time - but it's not that obvious that an explicit journal replay is
required prior to beeing able to mount the device.

Best Regards

----------
Linux l1 2.4.20

l1:/mnt # mount /dev/sda1
(Continue reading)

Permalink | Reply | 1 comment |
headers
webmaster | 10 Nov 2003 03:30

JFS chosen by Linisys for use with Shark Linux

Greetings from the Shark Linux Team!

Linisys, LLC of New Orleans Louisiana is the developer of a new distribution of
the Linux operating system. Shark Linux 1.05, scheduled for release before the
end of the 2003 has chosen IBM's JFS (Journaled File System) for its default
filesystem. 

Shark Linux is currently still in building stages (as of this writing), and is
based primarily from the gentoo-linux project.  The goal is to provide a
quickly deployable, optimized system, with pre-configured applications.  Shark
Linux aims to re-define the ANSI console by providing easy to use
administrative and userland tools for working within the system, therefore
eliminating the need for resource hungry, X-based GUI tools. A desktop version
will of-course utilize KDE and remote administration tools.

Although many of the tools will not be available in the first release, this will
be a great chance to test the deployment of JFS as a native filesystem for a
distribution.

-- Shark Linux Team
webmaster <at> sharklinux.com

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.
Permalink | Reply | 0 comments |
headers
Dave Kleikamp | 10 Nov 2003 16:37
Picon
Favicon

Re: Report: mounting JFS fails unless log got replayed

On Fri, 2003-11-07 at 20:32, mt <at> amira.ch wrote:
> Hello all,
> 
> as reported earlier on this mailinglist I had problems
> mounting an JFS partition - it failed to mount when
> connected to an external device.
> 
> Meanwhile I managed to make the problem reproduce & fixable:
> mounting the JFS partition fails when the partition's journal
> was not already replayed by fsck.

It should only fail to mount if the partition was not umounted cleanly. 
If the partition was unmounted the last time it was accessed, there
should be no need to replay the journal.  Is this the case?

> here the related lines from /etc/fstab
> #/dev/hdg1      /mnt/sda1       jfs     defaults                1 2
> #/dev/hdg2      /mnt/sda2       jfs     defaults                1 2
> /dev/sda1       /mnt/sda1       jfs     defaults,noauto         0 0
> /dev/sda2       /mnt/sda2       jfs     defaults,noauto         0 0
> 
> Note: I'm using an external IDE-Firewire bridge, so it's an SCSI-device
> to be mounted - anyway the IDE-HD is still the same.
> It's obvious that the on-boot scanning and error-attempt values have to
> be zero for an externally connected device that may be plugged in
> at any time - but it's not that obvious that an explicit journal replay is
> required prior to beeing able to mount the device.

Does it cause a problem if the last value in /etc/fstab is 2 and the
devices are not connected during boot?  If there are no ill effects, it
(Continue reading)

Permalink | Reply | 0 comments |
headers
kevin-jfs | 15 Nov 2003 15:42

tons of 'jfs_clear_inode, active_ag =' ... messages

Hi,

I am using JFS with 2.4.20 on debian, running on a PowerMac clone.  The
kernel is debian's 2.4.20-4-powerpc.  Is there a way to for me to get an
idea about what version of JFS kernel code I am actually running without
source code to my kernel?  I ran strings and nm on the JFS kernel module, and
didn't see anything helpful..

I can not seem to find the source package diffs for 2.4.20-4 (the -powerpc
diffs are seperate and appear to do nothing related to any filesystem code.)

Anyway, the messages I saw are here:

	http://blue.netnation.com/~nop/jfs-kernel-log.txt

All of the 'jfs_clear_inode, active_ag = ' messages appeared when I
was doing something along the lines of this:

 /jfs_partition/path/blah# tar cvf - /etc /lib /usr/lib /bin /sbin /usr/sbin
 /usr/X11R6 /usr/bin | tar xf -

The sources of the copy were all on ext3.  (I was trying to copy a bunch
of stuff to use in a chroot, to test 'prelink' out.)  A lot of files
with the same names existed in the target path, so tar would have been
doing some unlink()ing..

After seeing the huge amount of kernel messages, I suspended tar and
tried searching for the message in google..  Then I stopped tar,
unmounted the partition, at which point the 'dbUnmount' messages
appeared, and ran jfs_fsck with -n -v, which reported no errors
(Continue reading)

Permalink | Reply | 1 comment |
headers
szonyi calin | 17 Nov 2003 10:36
Picon
Favicon

Cannot delete file on jfs filesystem

Hi
There is a file in the kbd package with an unicode filename
which i cannot delete.
Kernel 2.6.0-test9 
jfs-utils 1.1.4
util-linux 2.11.z

ls says: No such file or directory
Default NLS for kernel is iso8859-2
utf is copiled in kernel 

Any ideea what can i do ?
I tried using the filesystem editor for jfs but it only had an 
"Modify inode" command

Thanks
Bye
Calin

=====
--
A mouse is a device used to point at 
the xterm you want to type in.
Kim Alm on a.s.r.

___________________________________________________________
Do You Yahoo!? -- Une adresse  <at> yahoo.fr gratuite et en français !
Yahoo! Mail : http://fr.mail.yahoo.com
Permalink | Reply | 1 comment |
headers
Dave Kleikamp | 17 Nov 2003 15:40
Picon
Favicon

Re: Cannot delete file on jfs filesystem

On Mon, 2003-11-17 at 03:36, szonyi calin wrote:
> Hi
> There is a file in the kbd package with an unicode filename
> which i cannot delete.
> Kernel 2.6.0-test9 
> jfs-utils 1.1.4
> util-linux 2.11.z

Which file is this?

> 
> ls says: No such file or directory
> Default NLS for kernel is iso8859-2
> utf is copiled in kernel 
> 
> Any ideea what can i do ?

You can try remounting the file system with "-i iocharset=utf8".

> I tried using the filesystem editor for jfs but it only had an 
> "Modify inode" command

If that doesn't work, you can run jfs_debugfs against the file system,
and use the dtree command against the directory containing the file. 
You'll need the inode number of the directory (ls -i).  You can probably
figure out how to navigate through the directory.

> 
> Thanks
> Bye
(Continue reading)

Permalink | Reply | 0 comments |
headers
Dave Kleikamp | 17 Nov 2003 16:02
Picon
Favicon

Re: tons of 'jfs_clear_inode, active_ag =' ... messages

On Sat, 2003-11-15 at 08:42, kevin-jfs <at> raindrop.net wrote:
> Hi,
> 
> I am using JFS with 2.4.20 on debian, running on a PowerMac clone.  The
> kernel is debian's 2.4.20-4-powerpc.  Is there a way to for me to get an
> idea about what version of JFS kernel code I am actually running without
> source code to my kernel?  I ran strings and nm on the JFS kernel module, and
> didn't see anything helpful..

Without the source, I wouldn't know.

> I can not seem to find the source package diffs for 2.4.20-4 (the -powerpc
> diffs are seperate and appear to do nothing related to any filesystem code.)
> 
> Anyway, the messages I saw are here:
> 
> 	http://blue.netnation.com/~nop/jfs-kernel-log.txt

These warnings are harmless, but annoying.  This problem was fixed last
December.

> All of the 'jfs_clear_inode, active_ag = ' messages appeared when I
> was doing something along the lines of this:
> 
>  /jfs_partition/path/blah# tar cvf - /etc /lib /usr/lib /bin /sbin /usr/sbin
>  /usr/X11R6 /usr/bin | tar xf -
> 
> The sources of the copy were all on ext3.  (I was trying to copy a bunch
> of stuff to use in a chroot, to test 'prelink' out.)  A lot of files
> with the same names existed in the target path, so tar would have been
(Continue reading)

Permalink | Reply | 0 comments |
headers
Zhihui Zhang | 1 Dec 2003 23:52
Favicon

Some questions of IBM's JFS


Hi,

I have been studying JFS on and off for quite a while. But I can never be
as sure as the experts who really work on JFS on a daily basis.  So would
you please clarify the following points for me:

(1) Dirty metapages are usually marked as nohomeok.  But metapages for
bitmap seem to be an exception.  Why treat bitmap metapage differently?
My guess is because of the careful write strategy JFS is using.  If so,
why working bitmap is updated first during an allocation, while permanent
bitmap is updated first during a free?

I understand some metapages (control nodes in the disk space map) does not
need to be logged because their information can be reconstructued easily.
When are they updated normally?

(2) The directory index table seems to try to bridge the legacy linear
offset and the real cookie in JFS - a leaf node number plus a slot number.
What if an entry's location changes after its index is put into the table
(e.g., a split of leaf node)?  Should there be something like
update_index() in addition to add_index() and free_index()?

The size of the directoy is the size of the directory index table.  The
index is persistent because it is never reused. The table must grow
forever because freed index is not reused.  How is the performance loss
due to maintaing such a large table?  Maybe we can simply encode the leaf
node number and the slot number into one 64-bit cookie. That would get rid
of the table.  I must miss something?
(Continue reading)

Permalink | Reply | 1 comment |
headers
Suman Puthana | 1 Dec 2003 08:45

question regarding disk I/O performance(iowait) on 2.4.21 kernel

Hello,

Is anybody aware of any changes in the 2.4.21 kernel which affects disk I/O
badly?  We are noticing that the cpu taken up by iowait's is causing
significant degradation in disk performance.

We have an application in which there are multiple processes writing to disk
simultaneously, each process writing anywhere from 16 - 64 K Bytes at a time
and sleeping for about 50 ms depending on how long the write() call takes to
finish. It is important that the write() finishes within 50 ms for this
application. The file's are opened with the open() call with the standard
O_RDWR and O_CREAT flags.

On the 2.4.18 kernel, 25 of these processes take up about 15% of CPU on a
dual Xeon Pentium 4 2.4 GHz processor with 1 MB of memory for a total
throughput of about 7 MBps which is very good.  The write() calls typically
take less than 5 ms to complete and we sleep for the remaining 45 ms or so.

But on the same system when we installed the 2.4.21 kernel, these 25
processes take anywhere between 40 - 70 % of the CPU depending on how  much
CPU the iowait is taking up. The iowait part of it varies all the way from
10 - 60 %. (The iowait CPU is within 1% on the 2.4.18 kernel.).  What is
even worse - sometimes the write() calls take anywhere between 1 - 2
seconds( yes seconds!) to return, which degrades the performance of our
application server pretty badly.

 It happens on all kinds of storage devices so it's definitely not a
hardware problem.(We tested on the standard IDE disk all the way upto the
EMC clariion storage system).
(Continue reading)

Permalink | Reply | 2 comments |

Gmane