F-prot auto updates ...

[Gerry Doris]

> Raymond Dijkxhoorn wrote:
>> Hi!
>>
>>>> Ok, thanks. Then i know its not just me :)
>>
>>> I only noticed a backlog yesterday 20:00 (GMT +2).
>>>
>>> Checking my log I can't see anything strange at that time. I do see
>>> an entry "Jul  7 19:02:13 netlx014 F-Prot autoupdate[7265]: F-Prot
>>> successfully updated." So it took just over 2 minutes to update my
>>> systems.
>>>
>>> Going back in my logs I noticed it normally takes 10 to 15 seconds to
>>> update F-Prot. But I don't think 2 minutes is that much of a problem
>>> when I update F-Prot.
>>
>> A little later this was fixed. Most likely due to network or system
>> problems @f-prot. It was also depending on what machine in their load
>> balanced cluster you were ending. So i guess one or two boxes that
>> would not do the right things :)
>>
>> All seems fixed now indeed btw.
>>
>> Bye,
>> Raymond.
>
> Should there not be an timeout built into the update process to ensure
> this does not happen?
>
> Dean Plant

I just looked at the /usr/lib/MailScanner/f-prot-autoupdate script and I
believe there is a built in timeout.  Julian is using wget to download the
f-prot files and is allowing 3 tries.  If the download isn't successful he
calls a bailout routine and exits.

If the timeout is too much then I suppose the easiest way to reduce it
would be to lower/remove the retries.

Gerry