savecrash giuves error. - savecrash: No dump header

[RAC_1]

Test Server - A class running 64 bit 11.00.

I am using savecrash from command line to save core. /etc/rc.config.d/savecrash and /etc/rc.config.d/crashconf are OK.
crashconf -v is as follows.
CLASS PAGES INCLUDED IN DUMP DESCRIPTION
-------- ---------- ---------------- -------------------------------------
UNUSED 4871 no, by default unused pages
USERPG 33514 no, by default user process pages
BCACHE 9478 no, by default buffer cache pages
KCODE 1813 no, by default kernel code pages
USTACK 577 yes, by default user process stacks
FSDATA 126 yes, by default file system metadata
KDDATA 10755 yes, by default kernel dynamic data
KSDATA 4402 yes, by default kernel static data

Total pages on system: 65536
Total pages included in dump: 15860

DEVICE OFFSET(kB) SIZE (kB) LOGICAL VOL. NAME
------------ ---------- ---------- ------------ -------------------------
31:0x01f000 310112 524288 64:0x000002 /dev/vg00/lvol2
----------
524288

I used it twice and it generated core.o and core.1 in /var/adm/crash.

After that it no more generates the crash files in /var/adm/crash. I used savecrash -rv -Z, savecrash -rv -Z -D /dev/vg00/lvol2 -O "offset"

For all commands it gives error as follows-
savecrash: No dump header

Any idea on what is happening?

Thanks in advance.

[Steven E. Protter]

lvlnboot -v

Just wondering, perhaps we're out of space in the dump area(which is usually swap).

How is the space in /var looking? Close to full? I hope not.

SEP.

[RAC_1]

SEP,

Forgot to add, there is enough space in /var.
System has 256 MB memory.
lvlnboot -v o/p as follows.-
Boot Definitions for Volume Group /dev/vg00:
Physical Volumes belonging in Root Volume Group:
/dev/dsk/c1t15d0 (0/0/1/1.15.0) -- Boot Disk
Boot: lvol1 on: /dev/dsk/c1t15d0
Root: lvol3 on: /dev/dsk/c1t15d0
Swap: lvol2 on: /dev/dsk/c1t15d0
Dump: lvol2 on: /dev/dsk/c1t15d0, 0

Moreover it dumped the crash twice and after that giving the above mentioned error.

[RAC_1]

ANY TAKERS?

[Sridhar Bhaskarla]

Hi Rac,

Were you sure the system actually dumped?. Did you see any errors in /etc/rc.log in the savecrash section if so?.

You will not alway have luck if you try to savecrash once system is up and running if the dump device is the only swap device you have on the system. So, I will be more concerned about the errors in /etc/rc.log.


What is your swapinfo -t output now?.

From your crashconf -v output, you don't need more than 75MB for dump. If you have space, then you can actually define a seperate dump only device and disable savecrash in rc scripts. This way the dump will not be disturbed and you can save it at your leisure.

-Sri

[RAC_1]

Shridhar,

It dumped the crash twice and after that it is giving error savecrash: No dump header.

Tried specifying seperate dump device, but no luck. Do I need to reboot after specifying seperate dump device?

lvlnboot -v after seperate dump device.

Boot Definitions for Volume Group /dev/vg00:
Physical Volumes belonging in Root Volume Group:
/dev/dsk/c1t15d0 (0/0/1/1.15.0) -- Boot Disk
Boot: lvol1 on: /dev/dsk/c1t15d0
Root: lvol3 on: /dev/dsk/c1t15d0
Swap: lvol2 on: /dev/dsk/c1t15d0
Dump: lvol12 on: /dev/dsk/c1t15d0, 0

[Sridhar Bhaskarla]

Hi,

You don't need to reboot the box. Does your crashconf -v reflect the new one?.

Do you have any error in /etc/rc.log near the savecrash section?.

-Sri

[Sridhar Bhaskarla]

Hi (Again),

Few more tests for you.

1. When you Toc the system, do you see it actually dumping to the device listed in "crashconf -v" output?. Do you see it completing 100%? to the device 31:0x01f000 .

2. When the system comes up, does it stay for a while in the "savecrash" section?.

You probably may have to redo your lvlnboots. Run lvrmboots followed by lvlnboots on all the devices.

Make sure you have an ignite tape so you will be safe while troubleshooting.

-Sri

[Sundar_7]

Hi,

Did u reboot the system after saving the core for two times ?

Since your dump device is /dev/vg00/lvol2 I beleive ur dump device is the same as the swap device.

If u have rebooted the system after saving the crash, the swap system overwrites the dump information in /dev/vg00/lvol2.

Thanks,

Sundar.