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

[Xen-users] Any idea of what can be done about Xen seemming to only service one Domain at the time under high Disk I/O load

  • To: xen-users@xxxxxxxxxxxxxxxxxxx
  • From: Henry Pepper <henrypepr@xxxxxxxxx>
  • Date: Tue, 12 Oct 2010 11:33:30 +0200
  • Delivery-date: Tue, 12 Oct 2010 02:34:38 -0700
  • Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:date:message-id:subject:from:to:content-type; b=SO4ytcOsfwCcrBvqxlK2mMmQ9fNGM89/Zoy9THoWqgWNOLGiYooLC+pbUddZO3cy74 GBUqMc1BxnxN7H5TCgttjqW0KKC56elcw6S1r65Ge5rtZqzDDh4Ia9F5yxZaPFwYeNXb AWb9Bi6qxfr9AT6xK4oiuBuTD2R4aLXUDHuYs=
  • List-id: Xen user discussion <xen-users.lists.xensource.com>


When running a disk performance tool in multiple DomU simultaniously,
then it seems that one DomU get done before the next DomU can start.

It seems as if once disk I/O has started between a DomU and the
Storage, then the hypervisor will continue servicing that DomU until
the DomU has a pause in the data requests, before the hypervistor
switches to the next DomU.

HW: Sun x4170
Xen: 3.4.3
Dom0:  CentOS release 5.5 (Final)
            Linux c6 2.6.18-194.el5xen #1 SMP Fri Apr 2 15:34:40 EDT
2010 x86_64 x86_64 x86_64 GNU/Linux
DomU: Same as Dom0.

1) What could be the problem?
2) What could I possibly do to troubleshoot this?
3) Anything I can do to configure Xen differently?



I've been running filbench

Xen-users mailing list



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