[410] in Zephyr_Bugs
slow server
raeburn@ATHENA.MIT.EDU (raeburn@ATHENA.MIT.EDU)
Sat Sep 12 21:37:20 1992
For reasons lost in the mists of time (I think I was chasing some bug), I do
"zctl ret" near the end of my zwgc startup script. Frequently these days,
that gets a timeout when I start up zwgc on tsx-11.
When I investigate this, I usually find tsx-11 talking to arilinn. According
to zstat's "packets handled" data, arilinn is getting only a couple packets
or so per second.
Since only (some of) the machines with "zephyr-test" zcluster data are using
arilinn for Zephyr service, and I'm doing this investigating at times when
I'd expect low usage, low rate isn't surprising. But arilinn is a DECstation
of some flavor; it's got no excuse for timing out. Any clues as to whether
it's the network, the machine, or the server? (The server is optimized,
isn't it? Is it compiled for profiling?)
