1 |
Hi all, |
2 |
|
3 |
Not really directly related to Gentoo except that the workstation where it |
4 |
happens is a Gentoo machine :-) |
5 |
|
6 |
Our VPN at work is a Juniper Secure Connect. I tend to have many konsole tabs |
7 |
running ssh at one time, as well as IMAP connections to Exchange and several |
8 |
browsers tabs to intranet sites. Occasionally, and apparently triggered |
9 |
randomly[1], all existing ssh sessions freeze and never come back (even after |
10 |
several hours). The connections are still up and do not die on the remote end. |
11 |
Mail connections stay up and the browser tabs continue to work as well[2]. |
12 |
|
13 |
I know the ssh sessions are alive on the other end as I can immediately log in |
14 |
in another konsole tab and that works just fine. Even after several hours the |
15 |
first sshd process does not timeout (timeouts are enabled). So it would seem |
16 |
that it's the return path going tits up. My next step is going to be running a |
17 |
trace and tcpdump on one of those existing session next time it happens. |
18 |
|
19 |
Meanwhile, and this is my real question, how would I go about sensibly |
20 |
debugging this in a methodical manner? I'm going to have to get my security |
21 |
guys and routing guys to investigate but they will need real information to |
22 |
work with. Everything I can think of just reduces to throwing brown smelly |
23 |
stuff at the wall and hoping some of it sticks. |
24 |
|
25 |
|
26 |
|
27 |
[1] By randomly I mean sometimes 4 times a day, sometimes nothing for several |
28 |
days. I have not seen a correlation to anything I'm doing that might trigger |
29 |
it. |
30 |
[2] This is expected as HTTP is sessionless. |
31 |
|
32 |
-- |
33 |
alan dot mckinnon at gmail dot com |