Ronan Heffernan wrote:
> Mike Branda wrote:
>
>> on these same render farm boxes, when I first installed the OS, the ssh
>> on them went through the paces and was at a password prompt in about 1
>> second. now it takes closed to 20 seconds.
>>
> try adding an entry to /etc/hosts on the 'slow' machine that resolves
> back to the machine which is running the SSH client. I have had this
> kind of problem with services that do a reverse-lookup (e.g. because of
> tcp-wrappers) upon connect. Do you have an in-house DNS setup for your
> render farm?
I agree with Ronan, DNS problems are usually the culprit with ssh
"hangs" on connect like this. A cluster like this really deserves a
caching DNS server and a private zone for your internal network.
Alternatively, you may be running into system load issues. Do you "nice"
your render jobs? What is the system load when you're attempting to ssh
in? Is the machine IO or CPU bound?
If the process simply appears to "hang", you can always use "strace -p"
or "ltrace -p" against the running pid to see what system/library calls
may be causing the holdup.
-- - Ian C. Blenke - Director of Service Delivery <icblenke@nks.net> (This message bound by the following: http://www.nks.net/email_disclaimer.html) ----------------------------------------------------------------------- This list is provided as an unmoderated internet service by Networked Knowledge Systems (NKS). Views and opinions expressed in messages posted are those of the author and do not necessarily reflect the official policy or position of NKS or any of its employees.
This archive was generated by hypermail 2.1.3 : Fri Aug 01 2014 - 16:45:54 EDT