Hi,
I found the following in the logs of a server that crashed:
Mar 19 14:09:28 sc4 kernel: Assertion failure in ext3_free_blocks() at
balloc.c line 358: "bh->b_committed_data != NULL"
Mar 19 14:09:28 sc4 kernel: Unable to handle kernel NULL pointer
dereference at virtual address 00000000
Mar 19 14:09:28 sc4 kernel: current->tss.cr3 = 39b8b000, %%cr3 39b8b000
Mar 19 14:09:28 sc4 kernel: *pde = 00000000
Mar 19 14:09:28 sc4 kernel: Oops: 0002
Mar 19 14:09:28 sc4 kernel: CPU: 0
Mar 19 14:09:28 sc4 kernel: EIP: 0010:[ext3_free_blocks+616/772]
Mar 19 14:09:28 sc4 kernel: EFLAGS: 00010282
Mar 19 14:09:28 sc4 kernel: eax: 00000060 ebx: c009c600 ecx:
ffffffff edx: 0000003c
Mar 19 14:09:28 sc4 kernel: esi: 00000adc edi: 00000000 ebp:
d57773a0 esp: f972de58
Searching through the ext3 mailing list I found the following post which
had a similar experience:
http://www.redhat.com/mailing-lists/ext3-users/msg02666.html
However, there was no reply to his query.
The server in question runs squid which handles high volume of traffic
every day.
Debian 2.2 (potato)
kernel 2.2.19
ext3-0.0.7
1024 MB RAM
Pentium III 850 Mhz
squid 2.4.STABLE4
Any help is much appreciated. Btw, would changing the kernel to 2.4.x
help in this situation?
Rgds,
Meng Kuan