Linux.com

i have a setup with about 30 Suse Linux Enterpise Server 10 Service Pack 1 machines and i have a backup script that in "pseudo code" looks like: for server in 1 to 30 do ssh $server tar -cz $logdir && scp $logdir.tar.gz $central_log_repo & done the problem is while all the servers get the logs tarred, the return trip doesnt allows work. random servers log this error: ssh_exchange_identification: Connection closed by remote host where of course remote host is the $central_log_repo. there is no mention on $central_log_repo of any attempt by the offending server(s) in summary, on one run server5 could scp the tarfile to $central_log_repo and on another it might fail because of the ssh error. how do i fix this?

I'm by no means an expert but you could try using only one SSH connection per PC, as it stands you use one to ssh into a server and another to transfer a file. ssh $server 'tar c $logdir' > $logdir.tar you could pass it through gzip too to optain a tar.gz like you had before(too lazy to try that myself though) You might also wanna try waiting for one transfer to finish before starting up another... who knows, it might be multiple connections happening too fast that's messing up things for you.

[quote=Egyptian]i'd rather fix the cause rather than the symptom. it is obvious the script isnt the problem as seen by the fact that it also affects clusterssh usage. so if it is "multiple connections happening too fast that's messing up things" - i'd rather fix that but i cant find the setting anywhere, not in sshd_config (which uses the default that case with sles10 sp1) nor in net.core.ipv4 nor iptables. anywhere else?[/quote] I do not know. I'm by no means a network guru, but the command I gave you halves the ssh connections you need for this operation so it is the Right Thing(tm) to do anyway(you only need an outgoing connection from your central repository, an ssh connection is bidirectional). I can't do tests on this issue because I only have three PCs here at home to play with, but if you want to test if it's too many connections happening at the same time that's causing the problem just have the script sleep for a few seconds between the ssh calls.

Sounds like sshd_config MaxStartups issue. sshd defaults to maximum of 10 simultaneous sessions that have not yet authenticated. I imagine it is a security feature so someone can't open a large number of brute force sessions. Note that, as soon as a connection is authenticated, it no longer counts against the MaxStartups limit. Assuming your script works for around 10 but not many more than 10 servers, this could be the cause. You could try increasing the MaxStartups amount. Or stagger your connections. Search the sshd_config man page for MaxStartups. Google for terms ssh_exchange_identification and MaxStartups to see references. E.g., http://archive.netbsd.se/?ml=openssh-unix-dev&a=2007-10&t=5430083