Exactly, Colin.  That's why a robust design must take such issues into
account, precisely because they bite so seldom.

Thanks for the AGIS notes.  They are extremely useful.

   j.

On 12 August 2010 11:54, Colin Allinson <[email protected]> wrote:
> John P. Hartmann wrote :
>
>> On 12 August 2010 10:08, Colin Allinson <[email protected]> wrote:
>> > c)      Very occasionally 2 servers would be doing exactly the same
> thing
>> > at the same time so they would clash.
>
>> No easy solution here.  You'll have to put a lock on the file and then
>> figure out what to do when the lock is held, particularly when the
>> lock is held, but the program that took the lock crashed for one
>> reason or another (or [shudder] you lost power).
>
>>    j.
>
> Sure, I guess there are things that could be done to make a clash even
> less likely than it is now but, from my perspective, it is not really a
> problem working with what we have now. Armed with the understanding of
> what is happening it is not really that difficult to code round these
> issues.
>
> The main reason I mentioned it was that, with my very busy servers, it was
> happening once or twice a day. However, in a less intense environment, it
> may happen once in a blue moon and be difficult to diagnose.
>
> Colin Allinson
> VM Systems Support
> Amadeus Data Processing GmbH
>

Reply via email to