Re: assertion failed: stmem.cc:270: "0" in Squid 3

Hmmm...Squid is configured to leave a core, but didn't. Is there
anything I'm missing to make it dump core? The directory is owned by
the Squid process and read/writable. The squid.conf specifies a
location for dumps.

Robert Collins wrote:
> I'll review that code section.
>
> A core is always good - you can then get squid up again and review the
> state at leisure...
>
>
> Rob
>
> On Thu, 2003-05-08 at 19:19, Joe Cooper wrote:
>
>>Hi Robert (I only single you out Robert, because your initials are in
>>the assert that comes right before this one in the code ;-),
>>
>>With large objects and/or range requests, I'm seeing the following:
>>
>>2003/05/07 22:19:35| ClientSocketContext::getNextRangeOffset: http offset 0
>>2003/05/07 22:19:35| ClientSocketContext::getNextRangeOffset: http
>>offset 1186
>>2003/05/07 22:19:38| ClientSocketContext::getNextRangeOffset: http offset 0
>>2003/05/07 22:19:38| ClientSocketContext::getNextRangeOffset: http
>>offset 1186
>>2003/05/07 22:19:38| ClientSocketContext::getNextRangeOffset: http
>>offset 12288
>>2003/05/07 22:19:41| memCopy: could not find offset 261 in memory.
>>2003/05/07 22:19:41| assertion failed: stmem.cc:270: "0"
>>
>>
>>The range request errors may be unrelated, since they happened a few
>>seconds later, but that's all that's in the logs before the crash. I
>>see the same crash everytime the Squid restarted after this point.
>>
>>In fact, it even looks kind of like a continuation of the same exact
>>request causing the error because it does this on the next restart/crash:
>>
>>2003/05/07 22:19:46| memCopy: could not find offset 262 in memory.
>>2003/05/07 22:19:46| assertion failed: stmem.cc:270: "0"
>>
>>And this one repeats a bunch (with restart jabber in between).
>>
>>I can probably get a gdb backtrace of this if needed. I'll have to wait
>>until tomorrow to put it back into a working environment though.

-- 
Joe Cooper <joe@swelltech.com>
Web caching appliances and support.
http://www.swelltech.com