On Sat, Oct 09, 1999 at 09:04:53AM -0500, Kevin Sawyer wrote:
> OK, I'm finally on to something.  I'm seeing this in my logs:
> 
> warning: trouble opening local/xx/xxxxx; will try again later
> warning: trouble opening remote/xx/xxxxx; will try again later

(This probably isn't very useful but anyhow...)

On Fridaymorning a system that handles diagnostic messages was being flooded
by thousands of emails coming from another system that had been having a
connectivity problem. The todo queue grew to over 10.000 entries and I was
seeing loads of the above messages as well. Scary.

Another thing I was seeing was that even after the todo queue had been emptied
and with >6000 messages still to be delivered locally, local concurrency kept
sticking around 1-2 out of 10.  When sending an ALRM to qmail-send this number
would temporarily go up to 10 but dwindle to 1 again shortly after.

Unfortunately, it turned out that because of the amount of logging output all
this activity produced, the interesting entries were pushed out of the logs
:-( I have doubled the queue depth since, so maybe next time I'll have some
hard evidence.

The thing that scared me most was that it took qmail over 15 minutes to
process the todo queue. As a result, several time-critical heartbeat messages
from other systems were not delivered in time, triggering (admittedly false)
alerts.

So, I'm really hoping that the new zeroseek technology that qmail 2 supposedly
will be using will address this todo issue.

-- 
Jos Backus                          _/ _/_/_/  "Reliability means never
                                   _/ _/   _/   having to say you're sorry."
                                  _/ _/_/_/             -- D. J. Bernstein
                             _/  _/ _/    _/
[EMAIL PROTECTED]  _/_/  _/_/_/      use Std::Disclaimer;

Reply via email to