graphite-dev team mailing list archive

[bhardy <question197816@xxxxxxxxxxxxxxxxxxxxx>]

New question #197816 on Graphite:
https://answers.launchpad.net/graphite/+question/197816

I am attempting to resize thousands of metrics and I am curious if anyone knows the most efficient way to do this (notes below).  The servers are beefy and attached to SAN.  

I shutdown carbon while resizing metrics.  If old find/exec is the only way to do this and will run for awhile, can this be done while carbon is running and not create any problems.  I use a top level directory to find all metrics.  Maybe going down a level or two and starting multiple resize scripts would also work to speed things up???


XARGS does not work as it attempts to pass more than one wsp file name to the resize script and I get an error.
find . -name *.wsp | xargs /usr/bin/whisper-resize.py 60:131400 --nobackup

OLD EXEC works but takes a long time due to it going thru each and every whisper file one-by-one and not doing so in bulk.
find . -name *.wsp -exec /usr/bin/whisper-resize.py {} 60:131400 --nobackup \;

NEW EXEC does not work and suffers from the same problem as xargs.
find . -name *.wsp -exec /usr/bin/whisper-resize.py 60:131400 --nobackup {} +

GNU PARALLEL works but takes somewhat as long as old exec.
find . -name \*.wsp | /usr/bin/parallel -j500 /usr/bin/whisper-resize.py {} 60:131400 --nobackup

-- 
You received this question notification because you are a member of
graphite-dev, which is an answer contact for Graphite.