[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-users] tracking down the Xen+NFS slowness problem


  • To: xen-users@xxxxxxxxxxxxxxxxxxx
  • From: "Mark Stier" <markstier@xxxxxxxxx>
  • Date: Wed, 23 May 2007 19:56:25 +0200
  • Delivery-date: Thu, 24 May 2007 10:07:12 -0700
  • Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:mime-version:content-type:content-transfer-encoding:content-disposition; b=EOZ2z+CEvbgBE/fgbHEymd8k/UPcWN9q/3ovjdAPtSCMtqVaRfL8mxHKwqtpal9IjWynfH+WfOYTW+bcAsNCcD0M9LBs8SzVlzG0857chPjk7SoyWL246LxQuzMPXYC3z4YldVC2vp86NUQPB7+SuSMaPTA/6BC2Er/JPCJtrlQ=
  • List-id: Xen user discussion <xen-users.lists.xensource.com>

Hello,

I recently had some problems with Xen timer and tcp congestion
algorithms, especially tcp_vegas which crashes machines because it
seems to measure round trip times to be zero -- and then divides by
them. So I thought I play around with congestion algorithms a bit in
order to if they have an effect on NFS throughput.

In the end, I detected a totally different correlation:

I have mounted an NFS share within domU. The share is exported from
dom0 on the same physical machine. I have a routed/masq setup. Writing
to the NFS share happens at approx 11 MB/s (much less than phsically
possible), reading happens at 1.3 MB/s. dom0 and domU are not running
*anything* else. The CPU usage is at approx 0% inside dom0 and domU.
Now what I found out: the read speed goes up dramatically when I start
bzip2 in dom0!

Best regards,
Mark

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.