|
|
|
|
|
|
|
|
|
|
xen-users
[Xen-users] vlans directly configured inside each domU using the default
Hello,
vlans directly configured inside each domU using the default xen bridge
have poor performance ...
I've configured somes vlan using 2nd technic that Felipe Alfaro explain
in it's blog :
http://felipe-alfaro.org/blog/category/xen/
For example with OpenLdap I have got very very poor performance
more than 3 minuts to make an ldap search on a small branch with less
250 entries ...
and this process finish to hang the domU ...
So I've look at /proc/interupts. I've read somewhere that a too high
value might significate a bridge configuration problem ... ???
#cat /proc/interrupts
CPU0
256: 39881 Dynamic-irq timer0
257: 0 Dynamic-irq resched0
258: 0 Dynamic-irq callfunc0
259: 235 Dynamic-irq xenbus
260: 3130 Dynamic-irq xencons
261: 5686 Dynamic-irq blkif
262: 1 Dynamic-irq blkif
263: 31658 Dynamic-irq eth0
I've setup this VM inside using one bridge per vlan with vlan configured
into dom0 ...
Now ldap search are very quick and the server seems to be stable
So all to seems works well except some problem with ARP
I've forced the MAC but sometimes seems that arp values are incorrect
... very strange !
I've got to 2 redundants routers with keepalived to going out to
Internet ...
So my default gateway is always a virtual IP ... and this seems to be a
problem with arp
Moreover value of eth0 interupts is + or - the half of the old config ...
#cat /proc/interrupts
CPU0
256: 25818 Dynamic-irq timer0
257: 0 Dynamic-irq resched0
258: 0 Dynamic-irq callfunc0
259: 237 Dynamic-irq xenbus
260: 819 Dynamic-irq xencons
261: 3434 Dynamic-irq blkif
262: 1 Dynamic-irq blkif
263: 16754 Dynamic-irq eth0
_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users
|
<Prev in Thread] |
Current Thread |
[Next in Thread> |
- [Xen-users] vlans directly configured inside each domU using the default xen bridge have very poor performance ...,
Sébastien CRAMATTE <=
|
|
|
|
|