Hi list,
I've got the "Mesg read: Connection reset by peer" problem
I see in the dumper.debug file on the server this line
dumper: time 27813.730: (sockaddr_in6 *)0x904c30 = { 10, 10080,
::ffff:145.238.170.197 }
dumper: time 27873.720: security_seterror(handle=0x8c66c18,
driver=0x8e2be0 (BSD) error=timeout waiting for REP)
dumper: time 27873.720: security_close(handle=0x8c66c18,
driver=0x8e2be0 (BSD))
and in the sendbackup.debug client
1223337320.985634: sendbackup: gnutar: /usr/libexec/amanda/runtar: pid
320
1223337320.985688: sendbackup: Started backup
1223339253.419941: sendbackup: critical (fatal): index tee cannot
write [Broken pipe]
In the wiki we can use "/proc/sys/net/ipv4/tcp_keepalive_time"
parameter but I don't find it for ipv6.
What i've unsterstood is that my tar of the dle on the client make the
computer too busy and he is loosing the connection.
( but the two another dle entry on the same client works great /home
and /etc )
What can i do ? the DLE is 700 G used for 1 T ( many files ). the
computer seems to be busy ( some physics computing ) when the backup
start
this DLE is nfs exported. I read before that's important to said.
Should i stay on ipv6 but change the dump style ( tar -> dump)
try connect with ipv4 and change the keepalive or change the security
method ( auth=bsd -> auth=tcpbsd )
thanks for helping me
julien
