Prev: Copied files take up more space than originals
Next: USB Drive: Can't open /dev/rdsk/c2t0d0s2: I/O error
From: underh20 on 22 Dec 2009 14:57
We have two Solaris 10 servers (A & B) setup for sending files between
them via SCP.
The keys/hosts configuration are set up between the two servers.
There's no
problem when sending individual file between the two servers.

However, when we are standing multiple files "cocncurrently" via shell
script or commands
from server A to server B, we are getting the following messages
repeatedly.
Some files got sent OK and some didn't go thru when the messages below
appear, regardless of
the size or type of these "unsent" files in server A :

:
:
ssh_exchange_identification: Connection closed by remote host
lost connection
ssh_exchange_identification: Connection closed by remote host
ssh_exchange_identification: Connection closed by remote host
lost connection
ssh_exchange_identification: Connection closed by remote host
lost connection
lost connection
ssh_exchange_identification: Connection closed by remote host
lost connection
ssh_exchange_identification: Connection closed by remote host
lost connection
:
:


Any idea how I could trouble-shoot and resolve this issue ?

Thanks,

Bill
From: OldSchool on 22 Dec 2009 15:44
On Dec 22, 2:57 pm, underh20 <underh20.scubadiv...(a)gmail.com> wrote:
> We have two Solaris 10 servers (A & B) setup for sending files between
>
> There's no problem when sending individual file between the two servers.
>
> However, when we are standing multiple files "cocncurrently" via shell
> script or commands
>

The first questions I'd have are:

How / what commands are being issued to send 'multiple files'
concurrently?
How many are being sent?
When this occurs, is it always around the same number of files?

I sort of suspect that you've run into a limitation with pty's not
being available on the server you're trying to connect to, or
something similar
From: Greg Andrews on 22 Dec 2009 18:48
underh20 <underh20.scubadiving(a)gmail.com> writes:
>
>We have two Solaris 10 servers (A & B) setup for sending files between
>them via SCP. The keys/hosts configuration are set up between the two
>servers. There's no problem when sending individual file between the
>two servers.
>
>However, when we are standing multiple files "cocncurrently" via shell
>script or commands from server A to server B, we are getting the
>following messages repeatedly.
>
>ssh_exchange_identification: Connection closed by remote host lost connection
>

One potential source is the configuration of the sshd daemon on
the machine that's receiving the connections. From the man page
for for the sshd config file "sshd_config":

MaxStartups
Specifies the maximum number of concurrent unauthenticated
connections to the SSH daemon. Additional connections will
be dropped until authentication succeeds or the LoginGraceTime
expires for a connection. The default is 10.


As someone else mentioned, a shortage of ptys can also cause similar
symptoms (though I think the error message will be slightly different).

-Greg
--
Do NOT reply via e-mail.
Reply in the newsgroup.
From: OldSchool on 24 Dec 2009 12:05
FYI,

The OP sent me a copy of the script he's running to do this. There
are 30+ lines of

nohup scp <some_source_file> <some_dest_file> &

so, as noted above, he's probably exceeded MaxStartups, since they'd
all process at approx. the same time....
From: Barry Margolin on 24 Dec 2009 16:49
In article
<18a46a6f-8ba7-4e5c-a95b-07078cf8b6b4(a)k19g2000yqc.googlegroups.com>,
OldSchool <scott.myron(a)macys.com> wrote:

> FYI,
>
> The OP sent me a copy of the script he's running to do this. There
> are 30+ lines of
>
> nohup scp <some_source_file> <some_dest_file> &
>
> so, as noted above, he's probably exceeded MaxStartups, since they'd
> all process at approx. the same time....

If you're copying a bunch of files to a common destination directory, I
believe it's best to do

scp file1 file2 file3 ... dest

I expect scp will reuse the same SSH connection where possible, avoiding
starting up lots of connections.

In some cases running all the transfers in parallel may provide a
benefit. If there's high latency between the machines, individual
connections may have limited throughput, so you'll get higher combined
throughput with concurrent transfers.

--
Barry Margolin, barmar(a)alum.mit.edu
Arlington, MA
*** PLEASE post questions in newsgroups, not directly to me ***
*** PLEASE don't copy me on replies, I'll read them in the group ***
 | 
Pages: 1
Prev: Copied files take up more space than originals
Next: USB Drive: Can't open /dev/rdsk/c2t0d0s2: I/O error