Detailed information.

Gil Disatnik Jewnix at technohac.com
Fri May 11 05:33:24 EST 2001


(Mr. Markus Friedl asked me to mail to this address)
Hello there.
My problem is as follows:
I am using OpenSSH 2.5.1 (Can't upgrade for now, explain will be given 
later...) on AIX 4.1.5 (Can't upgrade as well...)
I have written a distribution script that distributes tgz files along with 
their installation script, the distribution script simply scp the file and 
then reconnects and invokes the install script.
While I am loading the distributing machine the returned code from a 
remotely invoked script SOMETIMES (happens randomly) returns me -1 (255) 
instead of the real returned code.(When I say "loading" I mean approx 9 scp 
running simultaneously and approx 9 ssh install sessions while doing that...)

I can't upgrade for now as I am working for the Israeli Air Force in a 
top-secret facility, it takes quite a while to insert a code from the 
Internet (Intensive code checking...)

I can't even give you the debug information... but, this is very simple.
by checking two outputs I have ( using -v -v -v) while 1 of them was 
alright and the 2nd wasn't, the ONLY difference is that the "bad" session 
doesn't have this line:
debug: client_input_channel_req: channel 0 rtype exit-status reply 0
I am not a programmer, but it seems as if the sshd doesn't return the 
status reply at all, or at least- the client due to the extreme system load 
is not "picking" this signal...

Again: using ssh1 (the commercial version of ssh1) it DIDN'T happen (the 
system was loaded even more sometimes...)

sending a debug information of the server could be a bit problematic as it 
means I have to run it on all machines (a few hundreds...) - still, it 
seems like a bug that accures due to a high system load... so the sshd is 
probobly okay (I connected them one by one doing only exit 0... they all 
came out fine... doing exit 0 while distributing got screwd sometimes...)

The sample Mr. Markus Friedl gave me (sh -c 'for i in 1 2 3 4 5 6; do ssh 
localhost exit $i; echo $?; done') is irrelevant as it works fine for me as 
well (as I said, when I am connecting one by one - everything works fine, 
but when I am opening approx 10 ssh connections while 10 scp are killing 
the system/cpu it does happen).

Please tell me what you think...
Thank you very very much!



Regards

Gil Disatnik
UNIX system administrator

GibsonLP at EFnet
_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_
"Windows NT has detected mouse movement, you MUST restart
your computer before the new settings will take effect, [ OK ]"
--------------------------------------------------------------------
Windows is a 32 bit patch to a 16 bit GUI based on a 8 bit operating
system, written for a 4 bit processor by a 2 bit company which can
not stand 1 bit of competition.
-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.mindrot.org/pipermail/openssh-unix-dev/attachments/20010510/21ef8d89/attachment.html 


More information about the openssh-unix-dev mailing list