I've got an intranet mirror of centos, been updating it periodically, recently my updates started failing... I've tried several different mirrors as sources, but each time, the update fails at some random point with this error...
rsync: read error: Connection timed out (145) rsync error: error in rsync protocol data stream (code 12) at io.c(614) [receiver=2.6.8] rsync: connection unexpectedly closed (1605001 bytes received so far) [generator] rsync error: error in rsync protocol data stream (code 12) at io.c(463) [generator=2.6.8]
per a bit of googling, it appears this may be a generic 2.6.8 error, as explained here... http://www.mail-archive.com/rsync@lists.samba.org/msg16205.html
for what its worth, my local mirror server is actually a Sun ultrasparc running Solaris 10, with its built in rsync 2.6.8, which appears to be the same version as is running on the centos mirror servers.
¿Are you sure that it's actually the main reason of this behavior? ¿Is the rsync script exactly the same when it worked fine?
Jordi Espasa Clofent wrote:
¿Are you sure that it's actually the main reason of this behavior? ¿Is the rsync script exactly the same when it worked fine?
yes. it works sometimes, it fails other times. right now its failing mostly.
#!/bin/sh rsync -azH --delete --exclude=ia64 --exclude=s390* --exclude=alpha \ --exclude=centos/2* --exclude=centos/3.1 --exclude=centos/3.2 --exclude=centos/3.3 \ --exclude=centos/3.4 --exclude=centos/3.5 --exclude=centos/3.6 --exclude=centos/3.7 \ --exclude=centos/4.1 --exclude=centos/4.2 --exclude=centos/4.3 \ altruistic.lbl.gov::mirrors/centos /zdata/mirror
since it started failing mostly, i've tried several different nearby mirrors (stanford, kernel.org, lbl.gov, etc).
yes. it works sometimes, it fails other times. right now its failing mostly.
#!/bin/sh rsync -azH --delete --exclude=ia64 --exclude=s390* --exclude=alpha \ --exclude=centos/2* --exclude=centos/3.1 --exclude=centos/3.2 --exclude=centos/3.3 \ --exclude=centos/3.4 --exclude=centos/3.5 --exclude=centos/3.6 --exclude=centos/3.7 \ --exclude=centos/4.1 --exclude=centos/4.2 --exclude=centos/4.3 \ altruistic.lbl.gov::mirrors/centos /zdata/mirror
It's curious.
1. And ¿what happens if you downgrade rsync client-side to a previous version which runs fine? Maybe it will be enough, at least to determine the error side origin. 2. If this is a registered error of this rsync version you can try the same script without using compress tag (-z), because the rsync CPU consumption is always high.
Jordi Espasa Clofent wrote:
yes. it works sometimes, it fails other times. right now its failing mostly.
#!/bin/sh rsync -azH --delete --exclude=ia64 --exclude=s390* --exclude=alpha \ --exclude=centos/2* --exclude=centos/3.1 --exclude=centos/3.2 --exclude=centos/3.3 \ --exclude=centos/3.4 --exclude=centos/3.5 --exclude=centos/3.6 --exclude=centos/3.7 \ --exclude=centos/4.1 --exclude=centos/4.2 --exclude=centos/4.3 \ altruistic.lbl.gov::mirrors/centos /zdata/mirror
It's curious.
- And ¿what happens if you downgrade rsync client-side to a previous
version which runs fine? Maybe it will be enough, at least to determine the error side origin. 2. If this is a registered error of this rsync version you can try the same script without using compress tag (-z), because the rsync CPU consumption is always high.
I'm actually using rsync 4.6.8 on Solaris 10 (Sparc) as my Sun v210 has the available storage for our local mirror.
its running right now, if it aborts again, I'll remove the z and try again.
On Mon, 16 Apr 2007, John R Pierce wrote:
żIs the rsync script exactly the same when it worked fine?
yes. it works sometimes, it fails other times. right now its failing mostly.
#!/bin/sh rsync -azH --delete --exclude=ia64 --exclude=s390* --exclude=alpha \
[...]
Try removing -z - it puts extra load on the rsync server, and most of the data is already compressed, so compression would not give much. I seem to remember I had such problem when using -z on large files. Perhaps rsync server hits some resource limit when serving large files with compression. With updates size growing, more large files are present and the problem may be more frequent.
Good luck,
Wojtek
John R Pierce wrote:
I've got an intranet mirror of centos, been updating it periodically, recently my updates started failing... I've tried several different mirrors as sources, but each time, the update fails at some random point with this error...
rsync: read error: Connection timed out (145) rsync error: error in rsync protocol data stream (code 12) at io.c(614) [receiver=2.6.8] rsync: connection unexpectedly closed (1605001 bytes received so far) [generator] rsync error: error in rsync protocol data stream (code 12) at io.c(463) [generator=2.6.8]
per a bit of googling, it appears this may be a generic 2.6.8 error, as explained here... http://www.mail-archive.com/rsync@lists.samba.org/msg16205.html
for what its worth, my local mirror server is actually a Sun ultrasparc running Solaris 10, with its built in rsync 2.6.8, which appears to be the same version as is running on the centos mirror servers.
What errors is the Sun logging?
John Summerfield wrote:
John R Pierce wrote:
rsync: read error: Connection timed out (145) rsync error: error in rsync protocol data stream (code 12) at io.c(614) [receiver=2.6.8] rsync: connection unexpectedly closed (1605001 bytes received so far) [generator] rsync error: error in rsync protocol data stream (code 12) at io.c(463) [generator=2.6.8]
for what its worth, my local mirror server is actually a Sun ultrasparc running Solaris 10, with its built in rsync 2.6.8, which appears to be the same version as is running on the centos mirror servers.
What errors is the Sun logging?
I meant the client that logged those errors is a Sun. I'm assuming the mirror server that I'm fetching from is a CentOS system, I don't honestly know... I've tried several SF Bay Area mirrors as the source, gotten the same results.
John R Pierce wrote:
John Summerfield wrote:
John R Pierce wrote:
rsync: read error: Connection timed out (145) rsync error: error in rsync protocol data stream (code 12) at io.c(614) [receiver=2.6.8] rsync: connection unexpectedly closed (1605001 bytes received so far) [generator] rsync error: error in rsync protocol data stream (code 12) at io.c(463) [generator=2.6.8]
for what its worth, my local mirror server is actually a Sun ultrasparc running Solaris 10, with its built in rsync 2.6.8, which appears to be the same version as is running on the centos mirror servers.
What errors is the Sun logging?
I meant the client that logged those errors is a Sun. I'm assuming the mirror server that I'm fetching from is a CentOS system, I don't honestly know... I've tried several SF Bay Area mirrors as the source, gotten the same results.
For expedience, export by nfs as rw, and run rsync on something else.
-
John R Pierce -
John Summerfield -
Jordi Espasa Clofent -
Maciej Zenczykowski -
Wojtek.Pilorz