Bug 116337
| Summary: | aborting due to signal 6 (abort) | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | Red Hat Enterprise Linux 3 | Reporter: | Dan Barber <dbarber> | ||||||
| Component: | squid | Assignee: | Martin Stransky <stransky> | ||||||
| Status: | CLOSED DUPLICATE | QA Contact: | |||||||
| Severity: | high | Docs Contact: | |||||||
| Priority: | high | ||||||||
| Version: | 3.0 | CC: | bob, mindaugas, poelstra | ||||||
| Target Milestone: | --- | ||||||||
| Target Release: | --- | ||||||||
| Hardware: | i386 | ||||||||
| OS: | Linux | ||||||||
| Whiteboard: | |||||||||
| Fixed In Version: | Doc Type: | Bug Fix | |||||||
| Doc Text: | Story Points: | --- | |||||||
| Clone Of: | Environment: | ||||||||
| Last Closed: | 2005-10-11 09:42:03 UTC | Type: | --- | ||||||
| Regression: | --- | Mount Type: | --- | ||||||
| Documentation: | --- | CRM: | |||||||
| Verified Versions: | Category: | --- | |||||||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||||
| Embargoed: | |||||||||
| Attachments: |
|
||||||||
|
Description
Dan Barber
2004-02-20 08:04:06 UTC
Can you attach your squid.conf file to this so I'll have a chance of debugging it? The "FATAL: Received Segment Violation...dying." certainly appears to be an error, but without the config file, I can't even try to replicate it. http_port 80 icp_port 3130 cache_peer webc1.nocout.xxxx.xxx sibling 80 3130 hierarchy_stoplist cgi-bin ? cache_mem 512 MB cache_swap_low 72 cache_swap_high 75 maximum_object_size 32768 KB minimum_object_size 0 KB cache_dir diskd /var/spool/squid 32768 32 1024 store_dir_select_algorithm round-robin cache_store_log none debug_options ALL,1 ftp_user ftpuser acl all src 0.0.0.0/0.0.0.0 acl manager proto cache_object acl localhost src 127.0.0.1/255.255.255.255 acl webc1.nocout.xxxx.xxx src 10.10.20.113/255.255.255.255 acl nocman src 10.10.20.120/255.255.255.255 acl snmppublic snmp_community pubinner acl Safe_ports port 80 21 443 563 70 210 1025-65535 http_access deny !Safe_ports http_access allow all icp_access allow webc1.nocout.xxxx.xxx icp_access deny all miss_access allow all cache_mgr webmaster cachemgr_passwd secret all cache_effective_user squid cache_effective_group squid visible_hostname webc2.ior.xxxx.xxx httpd_accel_host virtual httpd_accel_port 80 httpd_accel_with_proxy on httpd_accel_uses_host_header on dns_testnames infosec.xxxx.xxx spawar.xxxx.xxx logfile_rotate 5 buffered_logs on always_direct allow nocman snmp_port 3401 snmp_access allow snmppublic localhost snmp_access deny all memory_pools on log_icp_queries off ipcache_size 4192 fqdncache_size 4192 It would also be helpful if you could enable core file dumps and setuid cores. If you could attach the core here it would be helpful in debugging this. Thanks! Created attachment 98238 [details]
patch to try fix a few shortcommings...
Based on the last core that jason (woodward) sent me regarding
this bug, This patch should clear up a few things. The core file submitted
clearly indicated that the segfault occured inside logfilePrintf(). My guess
is that some function was calling this with an improperly constructed format
string, or an invalid requisite argument. I was unfortunately unable to find
the problem, but I was able to relieve the segfault by rewriting the
logfilePrintf function as a macro in this patch. I understand this is a
completely undesireable fix, but since this group is under a deadline, I
thought this might help track down the real problem, or at least patch it up
until a better fix is constructed.
Also, this patch corrects something else that I noticed while investigating
this problem. The assert macro is used quite a bit in this app, but since it
runs a child as a daemon, any error messages reported due to failed asserts get
lost. This patch also adds a wrapper around assert to copy an error message
into syslog anytime an assert fails.
While I'm writing here, can I ask why squid is built with -fPIE and -pie turned
on? This causes it to be reported as a shared object, and makes for difficult
debugging, since you have to manually load it at the appropriate address in the
debugger.
Created attachment 98692 [details]
allow syslogging to work
The attached patch seems to fix our syslog issues on top of the fix the last
patch gave us.
While testing one setup we also got this error message. While googling around I found suggestion that upgrading to STABLE4 may solve the problem: http://www.mail-archive.com/squid-users@squid- cache.org/msg08444.html By the way our test was performed on x86_64 dual Opteron system. I get this under RHEL3U5.
It started when I upgraded
from: squid-2.5.STABLE3-6.3E.8
to: squid-2.5.STABLE3-6.3E.13
I backed down to:
squid-2.5.STABLE3-6.3E.9
And the problem has cleared up.
Could you please create and attach a strack trace when squid crashes? How to is here - http://people.redhat.com/stransky/squid.html Due to PIE gdb can't read symbols from the debug package. If your squid crashes (and it isn't a problem with assertion) and you can't obtain a stack-trace there are new packages which aren't compiled with PIE: http://people.redhat.com/stransky/squid.html *** This bug has been marked as a duplicate of 165367 *** |