[squid-users] Error Question

[Alex Rousskov]

On 2024-06-11 18:09, Jonathan Lee wrote:
> When I run sysctl debug.kdb.panic=1 I get a crash report for pfsense in 
> var/crash should my path for core dumps use my swap drive too?

It is a pfsense-specific question that I do not know the answer for. 
Perhaps others do. However, you may be able to get an answer faster if 
you set coredump_dir in squid.conf to /var/crash, start Squid with that 
configuration, and then kill a running Squid worker with SIGABRT.


HTH,

Alex.


>> On Jun 11, 2024, at 14:42, Alex Rousskov wrote:
>>
>> On 2024-06-11 17:06, Jonathan Lee wrote:
>>
>>> I can’t locate the dump file for segmentation fault it never 
>>> generates one.
>>
>> I assume that you cannot locate the core dump file because your 
>> OS/environment is not configured to produce core dump files. Enabling 
>> core dumps is a sysadmin task that is mostly independent from Squid 
>> specifics. The FAQ I linked to earlier has some hints, but none are 
>> pfsense-specific. If others on the list do not tell you how to enable 
>> coredumps on pfsense, you may want to ask on pfsense or sysadmin forums.
>>
>> Alternatively, you can try starting Squid from gdb or attacking gdb to 
>> a running Squid kid process, but neither is trivial, especially if you 
>> are using SMP Squid. The same FAQ has some hints.
>>
>> BTW, to test whether core dumps are enabled in your environment, you 
>> do not need to wait for Squid to crash. Instead, you can send a 
>> SIGABRT signal (as "root" if needed) to any running test process and 
>> see whether it creates a core dump file when crashing.
>>
>>
>>> I am running cache it shows a swap file however it is not readable.
>>
>> There are many kinds of swap files, but the core file we need is 
>> probably not one of them.
>>
>>
>>> I fixed the other issues.
>>
>> Glad to hear that!
>>
>> Alex.
>>
>>
>>>> On Jun 11, 2024, at 14:00, Alex Rousskov 
>>>> <rousskov at measurement-factory.com> wrote:
>>>>
>>>> On 2024-06-11 14:46, Jonathan Lee wrote:
>>>>> 2024-05-16 14:10:23 [60780] loading dbfile 
>>>>> /var/db/squidGuard/Nick_Blocks/urls.db
>>>>> 2024/06/11 10:23:05 kid1| FATAL: Received Segment Violation...dying.
>>>>> 2024/06/11 10:23:25 kid1| Starting Squid Cache version 5.8 for 
>>>>> aarch64-portbld-freebsd14.0...
>>>>> 2024/06/11 10:23:25 kid1| Service Name: squid
>>>>> 2024-06-11 10:23:25 [9471] (squidGuard): can't write to logfile 
>>>>> /var/log/squidGuard/squidGuard.log
>>>>> 2024-06-11 10:23:25 [9471] New setting: logdir: /var/squidGuard/log
>>>>> 2024-06-11 10:23:25 [9471] New setting: dbhome: /var/db/squidGuard
>>>>> 2024-06-11 10:23:25 [9471] init domainlist 
>>>>> /var/db/squidGuard/blk_blacklists_adult/domains
>>>>> 2024-06-11 10:23:25 [9471] loading dbfile 
>>>>> /var/db/squidGuard/blk_blacklists_adult/domains.db
>>>>> 2024-06-11 10:23:25 [9471] init expressionlist 
>>>>> /var/db/squidGuard/blk_blacklists_adult/expressions
>>>>> There is my log file being blocked for some reason
>>>>
>>>> Just to avoid a misunderstanding: This mailing list thread is about 
>>>> the segmentation fault bug you have reported earlier. The above log 
>>>> is _not_ the requested backtrace that we need to triage that bug. If 
>>>> there is another problem you need help with, please start a new 
>>>> mailing list thread and detail _that_ problem there.
>>>>
>>>>
>>>> Thank you,
>>>>
>>>> Alex.
>>>>
>>>>
>>>>>> On Jun 11, 2024, at 11:24, Jonathan Lee <jonathanlee571 at gmail.com> 
>>>>>> wrote:
>>>>>>
>>>>>> thanks i have enabled
>>>>>>
>>>>>> coredump_dir /var/squid/logs
>>>>>>
>>>>>> I will submit a dump as soon as it occurs again
>>>>>>
>>>>>>> On Jun 11, 2024, at 11:17, Jonathan Lee 
>>>>>>> <jonathanlee571 at gmail.com> wrote:
>>>>>>>
>>>>>>> I have attempted to upgrade the program fails to recognize 
>>>>>>> ”DHParamas Key Size” and will no longer use my certificates and 
>>>>>>> shows many errors. I am kind of stuck on 5.8
>>>>>>>
>>>>>>> I do not know where the core dump would be located on pfSense let 
>>>>>>> me research this and get back to you.
>>>>>>>
>>>>>>>> On Jun 11, 2024, at 11:04, Alex Rousskov 
>>>>>>>> <rousskov at measurement-factory.com> wrote:
>>>>>>>>
>>>>>>>> On 2024-06-11 13:24, Jonathan Lee wrote:
>>>>>>>>> FATAL: Received Segment Violation...dying.
>>>>>>>>> Does any know how to fix this??
>>>>>>>>
>>>>>>>> Please post full backtrace from this failure:
>>>>>>>> https://wiki.squid-cache.org/SquidFaq/BugReporting#crashes-and-core-dumps
>>>>>>>>
>>>>>>>> The other information you have already provided may help, but 
>>>>>>>> without a usable stack trace, it is unlikely that somebody will 
>>>>>>>> guess what is going on with your Squid.
>>>>>>>>
>>>>>>>> Please note that you are running Squid v5 that is no longer 
>>>>>>>> supported by the Squid Project. You should upgrade to v6+. 
>>>>>>>> However, I do not know whether that upgrade is going to address 
>>>>>>>> the specific problem you are suffering from.
>>>>>>>>
>>>>>>>>
>>>>>>>> HTH,
>>>>>>>>
>>>>>>>> Alex.
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> squid-users mailing list
>>>>>>>> squid-users at lists.squid-cache.org
>>>>>>>> https://lists.squid-cache.org/listinfo/squid-users
>>
>