GNU parallel with rsync

问题:

I'm trying to run some instances of rsync in parallel using ssh with GNU parallel. The command I'm running is like this:

find /tmp/tempfolder -type f -name 'chunck.*' | sort | parallel --gnu -j 4 -v ssh -i access.pem user@server echo {}\; rsync -Havessh -auz -0 --files-from={} ./ user@server:/destination/path

/tmp/tempfolder contains files with the prefix chunck and they contain the actual file lists.

With this command, I got the 4 calls for rsync alright, but they take a while to start running and don't start all together and don't run in parallel.

What am I doing wrong?

回答1:

Are you sure the rsyncs are really not running in parallel ?
Checking with ps | grep rsync while the command is running will show which and how many rsyncs are actually running simultaneously.

By default, parallel holds printing output from each job until it's finished so that the different commands' output don't get all mixed up together:

--group  Group output. Output from each jobs is grouped together and is only printed when the command
         is finished. stderr (standard error) first followed by stdout (standard output). This takes
         some CPU time. In rare situations GNU parallel takes up lots of CPU time and if it is
         acceptable that the outputs from different commands are mixed together, then disabling
         grouping with -u can speedup GNU parallel by a factor of 10.

         --group is the default. Can be reversed with -u.

My guess is the rsyncs are actually running in parallel, but from the output it feels like they're running serial. -u option changes that.

--

For example with this cmd:

$ for i in 1 2 3 ; do echo a$i ; sleep 1 ; done
a1
a2
a3

By default in parallel we get no feedback until it's all done:

$ (echo a ; echo b ; echo c ) | parallel 'for i in 1 2 3 ; do echo {}$i ; sleep 1 ; done  ' 
a1
a2
a3
b1
b2
b3
c1
c2
c3

Whereas with -u stuff get printed right away:

$ (echo a ; echo b ; echo c ) | parallel -u 'for i in 1 2 3 ; do echo {}$i ; sleep 1 ; done  ' 
a1
b1
c1
a2
b2
c2
a3
b3
c3

In both cases it took 3s to run though so it's really running simultaneously...