batch mode hangs if dest is uptodate

Bill Wendin bill at nowsol.com
Mon Dec 22 03:58:55 GMT 2008


I have been using batch mode successfully.  I try this using cygwin on a 
PC with 3.0.4 version of rsync.  The two PackageOlder dirs are identical

cygwin> diff -r /home/bill/tmp/PackageB /home/bill/PackageB

cygwin> rsync -vv --only-write-batch=/Temp/changeAtoB.rsync -rptO -L 
--delete-delay devserver::RT/PackageB/ /home/bill/tmp/PackageA
opening tcp connection to devserver port 873
sending daemon args: --server --sender -vvLtpre.iLs --timeout=180 . 
RT/PackageB/

receiving incremental file list
delta-transmission enabled
Upgrader.dat is uptodate
App.dat
App/junko.txt
bootstrap_upgrader/upgrader.exe is uptodate
App/rsync-2.4.6.tar.orig
deleting rsync-3.0.4.tar

sent 7046 bytes  received 5675 bytes  25442.00 bytes/sec
total size is 1331258  speedup is 104.65 (BATCH ONLY)

cygwin> rsync --read-batch=/Temp/changeAtoB.rsync -rptO -L 
--delete-delay /home/bill/PackageA
Setting the --recurse (-r) option to match the batchfile.
receiving incremental file list
delta-transmission enabled
Upgrader.dat is uptodate
bootstrap_upgrader/upgrader.exe is uptodate
App.dat
App/junko.txt
App/rsync-2.4.6.tar.orig
deleting rsync-3.0.4.tar

sent 7046 bytes  received 5675 bytes  25442.00 bytes/sec
total size is 1331258  speedup is 104.65

My comment: this works and changes PackageA/ to PackageB/

cygwin> rsync --read-batch=/Temp/changeAtoB.rsync -rptO -L 
--delete-delay /home/bill/PackageA
Setting the --recurse (-r) option to match the batchfile.
receiving incremental file list
delta-transmission enabled
App.dat is uptodate
Upgrader.dat is uptodate
App/junko.txt is uptodate
App/rsync-2.4.6.tar.orig is uptodate
bootstrap_upgrader/upgrader.exe is uptodate

My comment: 2nd run hangs if all files are up-to-date.  I would expect 
it to exit gracefully if /home/bill/PackageA is uptodate.  To debug, I 
tried using just -a on batch creation and batch read and it has same 
behavior.

I have hardware that can be turned off asynchronously during an rsync. 
I found this defect because the hardware was turned off between rsync 
finishing the batch update of files and rsync calling exit_cleanup (so 
rsync was flagged by my master code as never exiting correctly).  My 2nd 
re-run hangs as stated above.  I can work around it but any objection to 
filing this as a defect?



More information about the rsync mailing list