On Wed, Nov 17, 2010 at 2:35 PM, Mick <michaelkintz...@gmail.com> wrote: > On Wednesday 17 November 2010 20:10:54 Yohan Pereira wrote: >> On Thursday 18 Nov 2010 1:17:09 am Alan McKinnon wrote: >> > Let's compare then. My system is 4 minutes to re-index everything from >> > scratch. How long does your take and how big is your filesystem? >> >> time updatedb >> real 1m35.163s >> user 0m0.815s >> sys 0m2.454s >> >> this was the first time i ran it after installation. i had uinstalled >> slocate prior to this. >> >> PRUNEPATHS="/tmp /var/tmp /root/.ccache /media/stuff/backup/gentoo32" >> >> the last folder is my 32bit chroot. >> and i got arround 225 gb of data. i tried running a few qurries .. seems to >> have indexed everything >> >> also after this i tried deleting /var/lib/mlocate/mlocate.db and then >> running updatedb and it hardly took a second. >> >> time updatedb >> >> real 0m0.367s >> user 0m0.193s >> sys 0m0.167s >> >> weird indeed. > > I can't be doing this right ... > > I removed /var/lib/mlocate/mlocate.db, the ran updatebd: > > # time -p updatedb > real 113.22 > user 0.62 > sys 8.00 > > Then removed it again and run it again: > > # time updatedb > > real 0m1.063s > user 0m0.162s > sys 0m0.896s > > Why is the second time so much faster? The size of the derived db was the > same on both occasions.
I guess caching like Volker said too. What happens if you do something like this twice: sync; sh -c "echo 3 > /proc/sys/vm/drop_caches"; time updatedb
