Thread (53 messages) flat view 53 messages, 8 authors, 2011-05-10

Re: Scalability of interface creation and deletion

From: Eric Dumazet <hidden>
Date: 2011-05-07 16:26:33

Le samedi 07 mai 2011 à 16:26 +0100, Alex Bligh a écrit :
On the current 8 core box I am testing, I see 280ms per interface
delete **even with only 10 interfaces**. I see 260ms with one
interface. I know doing lots of rcu sync stuff can be slow, but
260ms to remove one veth pair sounds like more than rcu sync going
on. It sounds like a sleep (though I may not have found the
right one). I see no CPU load.
Here, on 2.6.38 kernel (Ubuntu 11.04 provided, on my 2 core laptop)

# time rmmod dummy

real	0m0.111s
user	0m0.000s
sys	0m0.000s


This removed my two dummy0/dummy1 devices.

On another machine with a very recent kernel :
$ modprobe dummy numdummies=1
$ ifconfig dummy0 192.168.46.46 up
$ time rmmod dummy

real	0m0.032s
user	0m0.000s
sys	0m0.001s
$ uname -a
Linux svivoipvnx001 2.6.39-rc6-00097-g6ac1576-dirty #550 SMP Sat May 7
00:12:26 CEST 2011 i686 i686 i386 GNU/Linux


So 260ms is a bit too much, maybe you hit yet another bug.



Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help