dying children?

[Julian Field]

Drew Marshall wrote:
> On 1 Aug 2008, at 14:36, Julian Field wrote:
>
>>
>>
>> Drew Marshall wrote:
>>> On 1 Aug 2008, at 10:09, Julian Field wrote:
>>>>
>>>> Richard Siddall wrote:
>>>>> Julian Field wrote:
>>>>>> Someone else showed me a message that suffered the same problem a 
>>>>>> few weeks ago. Unfortunately I don't think there's anything I can 
>>>>>> do about it, sorry. It's to do with nesting in the HTML analysis 
>>>>>> code. Once it gets too nested up, Perl segfaults.
>>>>>>
>>>>>> Jules
>>>>>>
>>>>>
>>>>> Jules,
>>>>>
>>>>> Does that mean it's something like an out-of-memory error in one 
>>>>> of the CPAN modules?  Can we fix it by getting the module author 
>>>>> to handle excessive nesting?
>>>> The most likely culprit is HTML::Parser, but I have direct evidence 
>>>> to back that. I just know that it's in the HTML parsing where it 
>>>> falls over. Does HTML::Parser contain any non-Perl code?
>>>
>>>
>>> Jules
>>>
>>> I seem to get a number of, what I think are, these types of mail 
>>> that choke MS and hold the child process up until it times out. Is 
>>> there any way that a mail that causes this sort of time out to be 
>>> automatically quarantined? Perhaps by changing the scan time out 
>>> from a batch time out to a message time out?
>>>
>>> The problem that I see is that if a batch has 10 messages in it 
>>> (Often mainly Spam) 1 of the messages chokes spam scanning, the 
>>> whole batch times out and lets the all the other spam messages 
>>> through for delivery.
>> It's not a timeout issue. If it hits this, it brings the entire Perl 
>> system crashing down. Wrapping it in an eval and timeout may not 
>> help. It would certainly add more overhead.
>>
>> I am open to all suggestions though!
>
> Ahh, I see. Can I request this anyway? I am seeing a number of time 
> outs (Like upwards of 50 since midnight today!) from SpamAssassin 
> which I just can't diagnose. The problem message always seems to stop 
> at the point it starts running body checks (Or at least that's the 
> line it displays last before hanging). I have done all the usual stuff 
> but as I can't easily capture the problem messages as they are 
> delivered after the time out limit is hit, I am struggling to put them 
> somewhere for some expert help.
I've actually got a couple of things to go to this weekend, so may well 
not have much time for coding.

It may be the start of next week....

Jules

-- 
Julian Field MEng CITP CEng
www.MailScanner.info
Buy the MailScanner book at www.MailScanner.info/store

Need help customising MailScanner?
Contact me!
Need help fixing or optimising your systems?
Contact me!
Need help getting you started solving new requirements from your boss?
Contact me!

PGP footprint: EE81 D763 3DB0 0BFD E1DC 7222 11F6 5947 1415 B654


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.