rsync "hang"

jw schultz jw at pegasys.ws
Sat Dec 20 06:51:07 EST 2003


On Fri, Dec 19, 2003 at 01:14:21PM -0500, Brian Cuttler wrote:
> 
> Hello rsync list,
> 
> I am running a Solaris 8 system and have rsync 1.6.2 protocal 14 installed.
> We have updated (wed 17-dec) to the latest Solaris 8 recommended patch set.
> 
> When performing an rsync of two local UFS partitions rsync "hung", never
> completed, didn't exit. What I forced an exit via kill signal core dumped.
> 
> Config of disks is that 1) Mirror by Solistice Disksuite rsync to
>                            non-mirrored UFS partition.
>                         2) This non-mirrored UFS partition is later
>                            copied to a second non-mirrored partition.
> 
> It was during this second stage that rsync hung.
> 
> There was nothing particularly odd about the files being moved,
> not size nor content. They where in fact the unzip of the 8_Recommended
> patch set.
> 
> We have been coping one partition to another in this mannor for
> several months using this procedure (extracted from larger procedure).
> 
> echo "Copy /bk1-dbX files to /bk2-dbX"
> foreach DBX (db1 db2 db3)
>     set FDIR="/bk1-$DBX/oracle"
>     set TDIR="/bk2-$DBX/oracle"
> 
>     echo "Copy of $DBX files from bk1 to bk2"
>     echo $FDIR $TDIR
> #    rsync -rltv --delete -e ssh "$FDIR"  "$TDIR"
>     rsync -rltv --delete "$FDIR"  "$TDIR"
>     echo ""
>     echo ""
> end
> 
> This problem has only occured once but because the intent is to
> copy our database files (Oracle) and the script also performs a
> database shutdown and restart its something we felt we'd best
> meantion to the list.
> 
> The database didn't restart because the script didn't complete the
> rsync and processing is sequential. So its not just a matter of the
> rsync error, production was down.
> 
> Am I abusing some flag or is there a limit on number of files, size,
> something that I inadvertantly hit ?
> 
> I'm thinking its an anomaly, rsync successfully moved the same data 
> from disk  a->b, has been running very well for months.
> 
> We did find a memory error in /var/adm/messages - but the time stamp
> was off by many hours, so I'm thinking that is unrelated.
> 
> There are no reported disk error, disksuite reports that it is in
> good shape as well. All disks are local.
> 
> Thanks for your assistance, if more info is needed please ask.

Excellent bug report.

You don't say what sort of memory error you had so of course
i'll point my finger and say "hardware error" :)

Rsync 1.6.2 is so old i don't know when it was released.
1.6.4-1.6.7 appear to have been released in December 1997.
Our OLDNEWS file doesn't even go back that far.  There have
been so many fixes (some security) since then it isn't
funny.  

The current released version is 2.5.7 protocol 26.  2.6.0 is
to be released very soon.


-- 
________________________________________________________________
	J.W. Schultz            Pegasystems Technologies
	email address:		jw at pegasys.ws

		Remember Cernan and Schmitt



More information about the rsync mailing list