[squid-users] Error in cache.log : xstrdup : tried to dup a NULL pointer

Now I'm really out of resources. I've been reading and reading all of the
FAQ, Mailling archives, etc.

I really need you're help on this one. Maybe the answer is/was on the
mailling archive but I'm lost on this one.

Squid run's fine for a few hour and then restart with this message in the
cache.log :
               FATAL : tried to dup a NULL pointer !!
               Squid Cache (Version ....) : Terminated abnormaly.
               CPU Usage : 1035.060 seconds = 675.660 user + 359.400 sys
               Maximum Resident Size : 0 KB
               Page faults with physical i/o : 30113
               Memory usage for squid via mallinfo()
                                      total space in arena : 25140 KB
                                      ordinary blocks 24935 KB
                                      small blocks 0 KB
                                      holding blocks : 772 KB
                                      free small blocks 0 KB
                                      Total in use : 25707
KB 102% ---> Now I don't think that's a good thing but it's always above
100%
                                      Total free
205 KB 1%

I've been monitoring with MRTG and here what I've got witch I think is
strange : cacheSysPageFaults and cacheMemUsage are growing at about the
same frequency, thay only fall down when squid restart (abnormaly). Now
I've been reading that PageFault on Linux is weird so I don't know what to
think.

Also It seem like the squid process is alway growing. But the way I
understand I squid works I should have enought memory for my chache (1Gig).

Plus, I see a lot of idnsCheckQueue : Id ..... : giving up after 20 trieds
and 115.4 sec. , is this normal (as I understand thing it's only the DNS
that does not answer).

So here my specs :

My system is : Red Hat 7.0 (Linux 2.2.16-22)
                  Squid 2.3.STABLE3-ldap-auth (with the group ldap add-on
from : http://www.fatgut.org/squid/group_ldap_auth)

Server is : Compaq Pentium 133 RAID with 7.5 Gig and 256M RAM

In the squid.conf i've got :

     cache_mem 4 MB
     cache_swap_low 70
                cache_swap_hight 80
                cache_dir ufs /usr/local/squid/cache 1000 16 256 --->
should mean 1Gig for cache
     maximum_object_size 8096 KB
                replacement_policy LFUDA
                reference_age 1 week
                memory_pools off ---> also tried with "on"

I'm using squidGuard for blocking and LdapGroup for authentification.

Anyway, I'm requesting you're precious help !!!

p.s : I've got a lot of statistic if necessary + try on another server
(Celeron 333 Mhz IDE drive and same software) and have the same problem.
Received on Fri Jun 15 2001 - 13:23:45 MDT