(code 12) at io.c(177) error

Fri Jun 20 07:35:31 EST 2003

We are using rsync to backup files from a bunch of computers to a central
server using rsync in --daemon mode. I've got a problem and are looking 
for suggestions. 

I'm not a newbie to rsync and have checked the archives without getting
any useable conclusions.

Our old server works great - a redhat 7.2 / 2.4.18 machine. But it's 
getting full and there's no more updates for redhat 7.x/8.x after the end 
of the year, so we're moving away from that. All of our client computers 
(redhat of all versions, as well as suse 8.1 and 8.2) work well with this.
The old server is also at a remote location which I am trying to conserve 
bandwidth to, so it's got not as good a connection to everything else as 
the new server at our datacenter.

The clients backing up to a new computer has problems.
Some can do rsync backups repeatedly with no problems. Other clients 
backup once fine, and then after that I get a problem.

rsync: read error: Connection reset by peer
rsync error: error in rsync protocol data stream (code 12) at io.c(177)

If I remove the contents on the server and backup again, it works fine. 
Then of course, the next incremental backup will fail like above. If I 
control-C on the client and break out during a backup, it will fail when I 
restart the backup. I have removed the lockfile on the server with no 
success, rebooted, the server, tried different file systems on the server, 
restarted rsyncd, tried the rsyncd executable from the old server on the 
new server with no difference. Now I'm back with a fresh configure and 
compile of rsync, and things are misbehaving consistently.

What else might I try?

client end:
Red Hat Linux release 7.2, P200mmx, 64MB ram, 128MB swap. rsync uses 6MB 
ram when running before it crashes, all ext3. rsync 2.5.6 from the rsync 
site, plain old configure, make, make install. 3c509 ethernet over a 
11mbps 5.8ghz radio link that is rock solid. HP switches on both ends.

[root at bht /]# /usr/local/bin/rsync -avvW --delete --exclude-from=/etc/rsync.exclude / root at 10.0.0.5::bht.midcoast.com
opening tcp connection to 10.0.0.5 port 873
building file list ... 
expand file_list to 4000 bytes, did move
excluding directory dev because of pattern /dev
excluding directory proc because of pattern /proc
excluding directory tmp because of pattern /tmp
expand file_list to 8000 bytes, did move
expand file_list to 16000 bytes, did move
expand file_list to 32000 bytes, did move
expand file_list to 64000 bytes, did move
expand file_list to 128000 bytes, did move
expand file_list to 256000 bytes, did move
excluding directory mnt because of pattern /mnt
excluding directory updates because of pattern /updates
done
rsync: read error: Connection reset by peer
rsync error: error in rsync protocol data stream (code 12) at io.c(177)
[root at bht /]# 
[root at bht /]# /usr/local/bin/rsync --version
rsync  version 2.5.6  protocol version 26
Copyright (C) 1996-2002 by Andrew Tridgell and others
<http://rsync.samba.org/>
Capabilities: 64-bit files, socketpairs, hard links, symlinks, batchfiles, 
              IPv6, 64-bit system inums, 64-bit internal inums

rsync comes with ABSOLUTELY NO WARRANTY.  This is free software, and you
are welcome to redistribute it under certain conditions.  See the GNU
General Public Licence for details.
[root at bht /]# 
[root at bht /]# traceroute 10.0.0.5
traceroute to 10.0.0.5 (10.0.0.5), 30 hops max, 38 byte packets
 1  blackcoffee (10.0.1.6)  5.505 ms  5.020 ms  4.371 ms
 2  carafe (10.0.0.5)  6.236 ms  4.152 ms  9.002 ms
[root at bht /]#
[root at bht /]# ping -c 100 10.0.0.5 |tail -n 2
100 packets transmitted, 100 packets received, 0% packet loss
round-trip min/avg/max/mdev = 2.941/4.972/17.582/2.139 ms
[root at bht /]# 
[root at bht /]# cat /etc/rsync.exclude 
/tmp
/proc
/dev
/mnt
/updates

[root at bht /]# cat /proc/version 
Linux version 2.4.18 (root at test.midcoast.com) (gcc version 2.96 20000731 
(Red Hat Linux 7.1 2.96-98)) #1 Fri Mar 8 16:27:13 EST 2002

New rsync server:

Suse 8.2, athlon XP2100, 1GB ram, 1GB swap, data backuped onto both ext2 
and reiserfs 3.6.24 on hardware raid5 with same results. 3c905 100-baset 
ethernet, no ethernet layer errors, no packet loss.

carafe:/usr/local/bin # cat /proc/version 
Linux version 2.4.21-rc8 (root at test2) (gcc version 3.3 20030226 
(prerelease) (SuSE Linux)) #1 Thu Jun 12 14:42:14 EDT 2003

carafe:/usr/local/bin # /usr/local/bin/rsync --daemon
this starts rsync for the server.

carafe:/usr/local/bin # more /etc/rsyncd.conf 
log file = /var/log/rsyncd.log

[bht.midcoast.com]
path= /backup2/bht.midcoast.com/
use chroot = true
max connections = 1
lock file = /var/log/rsyncd.bht.midcoast.com.lockfile
read only = no
list = yes
uid = root
gid = root
ignore nonreadable
ignore errors
hosts deny = *
hosts allow = 10.0.1.5

carafe:/var/log # tail rsyncd.log -n 5000 |grep bht
2003/06/19 15:58:37 [12356] rsync to bht.midcoast.com from bht.midcoast.com (10.0.1.5)
carafe:/var/log # tail rsyncd.log -n 5000 |grep 12356
2003/06/19 15:58:37 [12356] rsync to bht.midcoast.com from bht.midcoast.com (10.0.1.5)

carafe:/var/log # cd /backup2/
carafe:/backup2 # ls -la
total 25
drwxr-xr-x    4 root     root         4096 Jun 19 13:48 .
drwxr-xr-x   23 root     root          536 Jun 19 06:39 ..
drwxr-xr-x   14 root     root         4096 Jun 19 13:15 bht.midcoast.com
drwx------    2 root     root        16384 Jun 19 13:46 lost+found

-- 
/*
Jason Philbrook   |   Midcoast Internet Solutions - Internet Access,
    KB1IOJ        |  Hosting, and TCP-IP Networks for Midcoast Maine
 http://f64.nu/   |             http://www.midcoast.com/
*/