[Integrit-users] Bug in integrit 2.00 minor3.02

[Raik L. <Rai...@GM...>]

I have made a check with integrit after I deleted many files (about 
30.000) on my Windows-"E"-partition. By this check, integrit chrashs.
Up to now integrit everytime runs fine on my System and never crashed.

This is the conf-file:
known=./dat/mnt-e.odb
current=./dat/mnt-e.ndb
root=/mnt/e
/mnt/e          sIMC
!/mnt/e/pagefile.sys
!/mnt/e/temp

The Old-Database-File ist about 7MB, the New-Database-File is about 
300kB (because the many deleted files). I thougt, that this could be 
the reason, but Ed L. Cashin don't think so.

Integrit crash's at the and of its task, comparing the new Database with 
the old one:
open("./dat/mnt-e.ndb", O_RDONLY|O_NONBLOCK|O_LARGEFILE) = 9
munmap(0x45ff0000, 1979645953)          = 0
fstat64(9, {st_mode=S_IFREG|0640, st_size=344566, ...}) = 0
mmap2(NULL, 344566, PROT_READ, MAP_SHARED, 9, 0) = 0xb7fab000
--- SIGSEGV (Segmentation fault) ---
+++ killed by SIGSEGV +++

I have noticed, that /proc/meminfo shows a increasing "dirty"-value, a 
few seconds before integrit crashs.
normal state:				Dirty:               4 kB
14 seconds before integrit crashs:	Dirty:              40 kB
13 Seconds until crash:			Dirty:             112 kB
6 seconds:				Dirty:             108 kB
5:					Dirty:             964 kB
4:					Dirty:            2328 kB
3:					Dirty:            2584 kB
2:					Dirty:            5528 kB
Crash:					Dirty:            8600 kB

I could reproduce the crash 2times. From these run is the strace and the 
meminfo. Now integrit run's fine, I can't reproduce the crash anymore.

My System: RedHat 8.0, Kernel 2.6.9 (kernel.org), gcc version 3.2.2; 
glibc-2.2.93-5.

My Question is: Has anybody seen this bug before and has an idea, what 
the reason is?!

Bye!

Raik