SA times out

[DAve]

Julian Field wrote:
> 
> 
> Kai Schaetzl wrote:
>> Julian Field wrote on Wed, 02 Apr 2008 16:45:54 +0100:
>>
>>  
>>>>>  but perhaps a feature request could be a
>>>>> CLI switch to specify the message ID so MS only scans the particular
>>>>> message(s) that you're interested in observing.
>>>>>           
>>>> Good idea. I'll take a look. Would a single ID do?
>>>>       
>>> All done. It will be in the next release.
>>>     
>>
>> Ahm, Julian, now that I have used the MS debugging feature a few times 
>> I think being able to grab a single ID may be nice, but not really 
>> helpful for a production machine. I have to disable at least MS if I 
>> want to debug (otherwise it would "steal" the queue files) and usually 
>> this is not done within a few seconds, but takes at least five minutes 
>> or more, maybe repeatedly. It would be nice if I could specify an 
>> alternative queue directory, so I can run a MailScanner instance in 
>> parallel to the production daemon and debug files from that directory 
>> while the normal sendmail/MS operation isn't affected. I think this 
>> would be much more helpful than specifying a certain ID.
>>   
> You can stop MailScanner completely, then restart the incoming sendmail 
> (or whatever MTA you use) so that you are providing email service to 
> your users. Then run MailScanner on the particular ID you want to test 
> it with. Then when you are happy, resume normal operation.
> Stop everything and start incoming MTA:
>    service MailScanner stop
>    service MailScanner startin
> Run it on 1 id:
>    MailScanner --debug --id=<message-id>
> Start up everything normally
>    service MailScanner restart
> 
> Should solve the problem for you. Saves me writing more code :-)

In my case, in the time it took to run debug four times I gained 400 
messages in the queue. I don't get much time to ponder the results. What 
I did this week was dump the output to file and then alternate which of 
the servers I stopped MS on so as to spread the downtime.

I am considering pushing a VMWare install up on the network and then 
installing roundhouse, just for testing with future upgrades. Which is 
arguably the smart option.

DAve


-- 
In 50 years, our descendants will look back on the early years
of the internet, and much like we now look back on men with
rockets on their back and feathers glued to their arms, marvel
that we had the intelligence to wipe the drool from our chins.