Quit signal received.

Some hints:

Is there firewall between server and client?

     Sometimes firewalls break connections.

Is there any server side tcp/ip tuning? (keepalive, connection state ... etc)

there is a firewall, the timeout period is 1 hour. firewall on the client is disabled.

I have these network pramaters on the server.

net.ipv4.tcp_max_syn_backlog = 8192
net.core.netdev_max_backlog = 8192
net.core.rmem_max = 16777216
net.core.wmem_max = 16777216
net.ipv4.tcp_rmem = 4096 87380 16777216
net.ipv4.tcp_wmem = 4096 65536 16777216
net.ipv4.tcp_keepalive_intvl = 30
net.ipv4.tcp_keepalive_probes = 8
net.ipv4.tcp_keepalive_time = 300
net.ivp4.tcp_fin_timeout = 60
sunrpc.tcp_alot_table_entries = 64

Hi there,

Have you tried to set the time out to "0"?

Please change the group properties, and set "Client Inactivity Time out" to "0"

Thank you.

Carlos.

I see those on rare occassions with win2k8 clients.  They are caused by VSS provider and I have't seen good solution yet.

I changed "Client Inactivity Time out" to "0", reboot the client. the backup job failed again on E drive.

RPC error: RPC send operation failed; errno = An existing connection was forcibly closed by the remote host.

Looks like the job failed on a 19GB file. I am going to exclude the file and try again tonight.

Hrvoje Crvelin wrote: I see those on rare occassions with win2k8 clients.  They are caused by VSS provider and I have't seen good solution yet.

I managed to address my problem by setting client paralleism to 1 on affected clients (previously it was 2).  The reason why 2 didn't work was due to HW VSS provider mishandling situation when VSS system saveset would be done/taken and at the same time snapshot for C drive which hosts this files was in process.  Thus I had to implement sequential execution for these clients.