SEO

Helo.

oslevel 5.3.0.0
machine: powerpc IBM,9110-510


[power64/admgk]$ /usr/sbin/backup -f /dev/rmt0 -0 -u /max
backup: The date of this level 0 backup is Tue May 9 12:00:10 CDT 2006.
backup: The date of the last level 0 backup is the epoch.
backup: Backing up /dev/rfslv03 (/max) to /dev/rmt0.
backup: 0511-251 The file system is still mounted; data may not be
consistent.
Use the umount command to unmount the filesystem; then do the
backup.
backup: Mapping regular files. This is Pass 1.
backup: Mapping directories. This is Pass 2.
backup: There are an estimated 20126938 1k blocks.
backup: Backing up directories. This is Pass 3.
backup: Backing up regular files. This is Pass 4.
backup: 3.86% of the backup is done. It should be finished in 124:27.
............
backup: 60.39% of the backup is done. It should be finished in 49:12.
backup: 0511-202 Internal error; received a segmentation violation.
backup: Child process 442398 returns status 206.
[power64/admgk]$ errpt
IDENTIFIER TIMESTAMP T C RESOURCE_NAME DESCRIPTION
B6048838 0509134006 P S SYSPROC SOFTWARE PROGRAM ABNORMALLY
TERMINATED

[power64/admgk]$ errpt -a

...............................
INODE NUMBER
2
PROCESSOR ID
0
CORE FILE NAME
/core
PROGRAM NAME
backbyinode
ADDITIONAL INFORMATION
raise 40
??

Symptom Data
REPORTABLE
1
INTERNAL ERROR
1
SYMPTOM CODE
PIDS/5765E6200 LVLS/520 PCSS/SPI2 FLDS/backbyino SIG/6 FLDS/raise VALU/40


I have used google but did not find anything usefull.

Any idea how to fix or workaround the problem

with regards
g.kalinski

Dzien Dobre (or something like that).

It is already in the first few lines: make the backup with an unmounted
filesystem ofr a level 0 backup !

What is the prupose of this backup ? You might want antoher method ..

Marcel

Bobohoolie wrote:

> Dzien Dobre (or something like that).
>
> It is already in the first few lines: make the backup with an unmounted
> filesystem ofr a level 0 backup !
>
> What is the prupose of this backup ? You might want antoher method ..
>
> Marcel
>
Hello.. (dzieñ dobry)

Just about month ago it works without any problem
(on mounted filesystem). I can use tar, but when
a file is larger than 2G it is not possible to restore it.
I have changed the backup procedure (use a tar),
and stop using files larger than 2G, but it takes
much more time.

dump is faster than tar.

But on this filesystem is about 210000 files.
Perhaps this is a cause.

with regards
g.kalinski


log from Apr 13 2006:

rmt0 Available
backup: The date of this level 0 backup is Thu Apr 13 23:19:11 CDT 2006.
backup: The date of the last level 0 backup is the epoch.
backup: Backing up /dev/rfslv03 (/max) to /dev/rmt0.
backup: 0511-251 The file system is still mounted; data may not be
consistent.
Use the umount command to unmount the filesystem; then do the
backup.
backup: Mapping regular files. This is Pass 1.
backup: Mapping directories. This is Pass 2.
backup: There are an estimated 18837747 1k blocks.
backup: Backing up directories. This is Pass 3.
backup: Backing up regular files. This is Pass 4.
backup: 3.68% of the backup is done. It should be finished in 130:49.
...............
backup: 99.72% of the backup is done. It should be finished in 0:18.
backup: There are 18826948 1k blocks on 1 volumes.
backup: There is a level 0 backup on Thu Apr 13 23:19:11 CDT 2006.
backup: The tape is rewinding.
backup: The backup is complete.
Fri Apr 14 01:26:21 CDT 2006

On 2006-05-09, Bobohoolie <juliorc@zonnet.nl> wrote:

> It is already in the first few lines: make the backup with an unmounted
> filesystem ofr a level 0 backup !

That is not an error, but a warning that the backup you're making might
very well be useless because the filesystem is still mounted. It should
not cause the program to get a segfault. It most likely means there's
a bug in backbyinode (used by backup). So, either upgrade to the latest
technology level/service pack, or contact IBM.


Note though that the warning is there for a reason, and the original
poster most likely should use backing up by file instead of inode,
given a mounted filesystem. As a bonus, this might even mean that
the bug in backbyinode doesn't get triggered.




--
Jurjen Oskam

Savage's Law of Expediency:
You want it bad, you'll get it bad.