Hi,
--- h/w && s/w ---
SPARC
Sol26
ssh 2.0.12
---
Have anyone noticed sshd to fail and drop connections on highly utilized
systems. I have a machine which acts as a jumphost into a production
system. About 80 ppl or more are always logged in working (active).
Sometimes there are one or two sshd processes that get stuck in a loop
and consume alot of CPU (~25% each). Those processes are killed off and
aren't considered that big of a problem (yet). A rather high percentage also get
their connections dropped
"Disconnected; connection lost (Connection closed.)."
during inactivity (all use shells that has no autologout
function). Yes I have tested if there is a correlation between the looping
sshd and the dropped connection and it is not.
Anyone seen a related scenario ?
rgds
/PL
