[arch-dev-public] [testing] mlocate replacing slocate

Andreas Radke a.radke at arcor.de
Sun Nov 4 03:29:19 EST 2007


Am Sat, 3 Nov 2007 17:48:51 +0100
schrieb Andreas Radke <a.radke at arcor.de>:

> I'm not sure how mlocate follows the kernel file syscalls and if it
> will slow down any file operations. I'm also concerned how it works in
> chroots. I might try it later.

Ok, have to say sorry here. Mixed mlocate up with rlocate :-/

Here are my experiences:

old slocate
[root at workstation64 andyrtr]# time updatedb

real    1m7.929s
user    0m1.683s
sys     0m2.566s
[root at workstation64 andyrtr]# time updatedb

real    1m19.958s
user    0m1.690s
sys     0m2.696s

new mlocate:
[root at workstation64 andyrtr]# time updatedb

real    10m33.805s
user    0m1.463s
sys     0m33.304s
[root at workstation64 andyrtr]# time updatedb

real    0m21.013s
user    0m0.617s
sys     0m1.333s
[root at workstation64 andyrtr]# time updatedb

real    0m1.183s
user    0m0.463s
sys     0m0.417s
[root at workstation64 andyrtr]# time updatedb

real    0m0.868s
user    0m0.413s
sys     0m0.443s


That'really faster after the first updatedb call! So I'm fine with it.

Andy




More information about the arch-dev-public mailing list