mksysb problem 2

[Guest_imported]

hello!

does anyone encountered similar problem as what i got while trying to backup system using mksysb?

i got this error (failed):

Creating......
Backing up....
172 of 32667 (0%)....../usr/bin/mksysb[801]: 33908 Bus error (coredump)

0512-005 mksysb: Backup Completed
The backup command completed with errors.
The message displayed on standard error contained additional information.

 

[aixqueen]

check errpt -a | more

and see if there are any hardware problems..
you also may end up going into service and fsck your filesystems ......

you can try and look at the core that was created.....also..... but I would check
smit.log and errpt to see if there is anything lurking.

 

[aixqueen]

check errpt as I said above...and look for hardware errors.....also:
You can try some of the below to try and pin point it...


Mksysb failed:
Try to send the mksysb to a file first

mkszfile && mksysb /dev/null 2>&1 /tmp/logfile

(this should come back quickly with an either ok or fail.
It will put errors in the logfile) Then do the following:

cd /usr/sbin/
cp mkinsttape mkinsttape.orig
vi mkinstalltape
add the line set -x at the top line of the file under the comments:

set -x (right above the line ) cleanup ()
Then save the file (esc wq!) then run the following:
#> mkinsttape /dev/null > /tmp/backup.log 2>&1
as root

Then when it is done look at the log file /tmp/backup.log
See what has been backed up…what was not…..any errors?….

put the original mkinsttape file back.

cp mkinsttape.orig mkinsttape

(This will tell you if the original required system files for a
bootable tape are there. If they are missing you have system
corruption, with possibly proto files or other required files missing.)

Two ways to run a verbose listing of what is being backed up in case it fails
on a file or link.
ksh -x mkszfile 2>&1 | tee -a /tmp/mkszfile.out
ksh -x mksysb /dev/rmt0 2>&1 | tee -a /tmp/mksysb.out

 

[Guest_imported]

hi aixqueen,

yeah, i tried to check using fsck and got these from the Filesystems (FS) indicated:

# fsck -p /dev/hd9var
/dev/hd9var (/var): Bad Inode Map (NOT SALVAGED)
/dev/hd9var (/var): Bad Block Map (NOT SALVAGED)
/dev/hd9var (/var): Filesystem integrity is not guaranteed
/dev/hd9var (/var): 424 files 52091 blocks 78981 free

also, same for /tmp

# fsck -p /dev/lv08
/dev/lv08 (/usr/b): Cannot read: block 360448 (NOT CLEARED) (TERMINATED)

are there any way/things to do to repair these FSs? aside from fsck?

i tried backing up using "savevg" for each two VG's but got the same error as the original post.

P.S. i tried running fsck but not in single mode, kinda hesitant to reboot the system coz of the thought it won't come up again or reboot.

 

[aixqueen]

fsck is the only way I know to really get it back.......however, if fsck cannot fix something it will delete it........(depending on the flags...) And hopefully you
have an old backup of some of the files...

You have to unmount something to get it to fsck....and neither var nor tmp do well except in service mode....to fsck....... (I suppose it can be done...but not easily and it is more trouble than it is worth...)

You could try and back up your stuff only....as in your database and your data....
in case something gets messed... (from a specific filesystem only...not all of var.....

http://techsupport.services.ibm.com...mode=9&documents=390605204914686&database=aix

------------------
One word of caution...if you reboot, you may get hung on 518...Here is the document
for the recovery....Have it printed out before you proceed...

http://techsupport.services.ibm.com...mode=9&documents=290605188914738&database=aix

---------------------------------------------

http://www.rs6000.ibm.com/doc_link/en_US/a_doc_lib/aixuser/usrosdev/check_filesys_integrity.htm

--------------------------------------

http://www.rs6000.ibm.com/doc_link/en_US/a_doc_lib/aixbman/baseadmn/damage_fix.htm

----------------------------------

http://www.rs6000.ibm.com/doc_link/en_US/a_doc_lib/cmds/aixcmds2/fsck.htm

---------------------------------------

 

[Guest_imported]

hi aixqueen,

thank you for the procedures, but i tried to boot up from cdrom:

* bootable cdrom
* choose maintanence mode for sys recovery
* access a root volume group (select 0)
* select rootvg by number
* access this vol group and start a shell before mounting FS (option2)
* finally got this error:

importing volume group...
rootvg
getlvodm : option requires an argument -- l
usage : getlvodm [-a LVdescript][......]
0516-104 lquerylv : incorrect LV identifier. the format must be VG_idetifier. LV_minor_number
could not find "/" and/or "/usr" FS
exiting to shell.

* df -k shows:

/dev/ram0 /
/dev/cd0 /SPOT

* initially, tried fsck -p for each FS and seems okay, then sync 3x; shutdown -Fr

* same problem exist (failed message after mksysb).

the system is still running though, is it okay if i backup the system on a per FS basis using tar? and re-install aix 4.2.1 again before restoring? maybe im just too panicky with this sytem.

 

[aixqueen]

Well that is distressing...you have corruption which will probably bite you eventually....
If I had an old mksysb that was good I would probably restore it...or do a preservation install....but I would make sure that my data at least was backed up before I did it...
Or I would try to move my stuff to another filesystem. Create a new filesystem ....move my stuff...then delete the old filesystem and put it back...maybe that will
fix some of the problems...you can't do that for the major ones like /var and /tmp...
you can also do an lppchk -v and lppchk -c and see if the system has broken filesets... Can you do any lsvg command or lspv command? If not you have ODM corruption as well.......

http://techsupport.services.ibm.com...mode=9&documents=290605223414650&database=aix

When the ODM entries for a volume group are corrupted, you may notice one of the following symptoms:

The lsvg, lslv or lspv commands fail.
You cannot change the size of file systems.
The system cannot find volume groups or device IDs.
The output of the following command contains question marks (?s) in some fields, or some fields are blank:
lsvg -l [VGname]

I am happy your system is up, but I would not have a warm and fuzzy about it...I would make sure I backed up at least your data...so if you do have to do a preservation install........or a restore from a mksysb your stuff is current...

Good Luck........the fsck may have removed some files that are needed because they were so corrupted.....Make sure you look in the errpt for hardware errors on the disk...that can cause problems as well........... A call to IBM may be a good thing to try and fix what problems you have...sounds like multiple ones....
my guess is that they will tell you to try and restore from mksysb.....

Again sorry your system is sick.

 

[Guest_imported]

hi aixquen,

thank you for the tips. yes! i think my system is sick since it started falling apart from sysboard, async card, power supply and hdisks.

my option now i guess is to replace the hd after backing up the system, re-instaled aix 4.2.1 os and restore FS one-by-one.

 

[vantandem]

Johny2001,

I’ve seen your error before:
mporting volume group...
rootvg
getlvodm : option requires an argument -- l
usage : getlvodm [-a LVdescript][......]
0516-104 lquerylv : incorrect LV identifier. the format must be VG_idetifier. LV_minor_number
could not find "/" and/or "/usr" FS

Sometimes you may get this error because you may have multiple internal scsi disks. If you do not choose the scsi disk that contains rootvg then you will get this error. When selecting the scsi drives it will display the hd5, hd9var, hd2, hd4 logical volumes. If you did see this, after you made your selection then you may have some corrupt data.

vantandem

 

[Guest_imported]

vantandem,

i only have 2 hdisks; hdisk0 (rootvg) & hdisk1 (datavg).

seems that /tmp & /var FSs are corrupted.