Well, I found out where all that load was coming from.
Looks like a recent Thunderbird update reset its config to "keep all
messages for this account on this computer" - for all accounts
Given that I'm the admin for some servers, and a bunch of email lists, I
keep filtered spam and viruses for analysis, and I keep email going back
about 30 years.... that kind of causes a lot of synchronization traffic
- to the point of really bogging down our imap daemon.
Consider this a friendly warning for those of you who use Thunderbird &
IMAP!
Cheers,
Miles Fidelman
Hi,
I wonder if anybody might have some thoughts on this:
Lately, our server has been showing high loading, but top shows that
the CPU is mostly in wait mode, and iotop shows low disk i/o traffic.
Usually, the system shows high load when either
a. Someone is indexing our web server (which shows up as CPU load on
apache).
b. Our list manager is processing a bunch of traffic (most of the load
comes from spam and virus filtering - which shows up as cpu load on
amavisd, and clamd, as well as a lot of disk i/o).
The one place I see this effect behavior, if for email clients
accessing IMAP - I've been seeing timeouts, and exceptionally high
delays to load an inbox, or move stuff to the trash.
This pattern is new - and there's really nothing that I've changed in
terms of configuration or loading that I can see.
So... any thoughts? Any diagnostic approaches?
Thanks much,
Miles Fidelman
--
In theory, there is no difference between theory and practice.
In practice, there is. .... Yogi Berra
