[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] Re: [Xen-devel] Limit I/O for individual disks of VM
If i run the exact same kernel without Xen and test the i/o, it won't crash.I just tried to look up an old post to refer it to You and seems it has been updated just yesterday with a fix: http://xen.1045712.n5.nabble.com/Xen-4-1-3ware-9690SA-rejecting-I-O-to-offline-device-td3208156.html Still, i will try to change the raid controller and see what happens then. Regards, Tõnis Bramanis On 28.09.2011 10:38, Joseph Glanville wrote: Hi, To add to the above it looks like you have a device driver or hardware issue with your harddrive or raid controller. I/O ratelimiting might hide the symptoms but not correct the problem. After you have addressed that you can look at implementing rate limiting via the methods described below: cgroups blk-io controller: You can find documentation etc in the linux git repo at /documentation/cgroups/blkio-controller.txt dm-ioband: Wiki is here http://sourceforge.net/apps/trac/ioband/wiki/dm-ioband Both of these are external to Xen for the reasons outlined by previous posts. Kind regards, Joseph. 2011/9/28 Jeremy Fitzhardinge <jeremy@xxxxxxxx <mailto:jeremy@xxxxxxxx>> On 09/27/2011 11:25 PM, Tõnis Bramanis wrote: > Hello > > and thank You for the answer. Has anyone of You tried to limit i/o on > the VMs? > > I have about 40 containers and when the i/o gets too high, the host > will crash with this error: > [23959.739698] sd 4:0:0:0: rejecting I/O to offline device > 5 [23959.771158] Buffer I/O error on device dm-0, logical block 40259 > > This is not a support request! I am just referring why i would like to > limit the disks i/o. I think that IO error is a real problem that you should try to solve, rather than work around it by rate-limiting IO. J > > Regards, > Tõnis Bramanis > > On 27.09.2011 22 <tel:27.09.2011%2022>:58, Jeremy Fitzhardinge wrote: >> On 09/27/2011 01:33 AM, Tõnis Bramanis wrote: >>> Hello! >>> >>> Will disk i/o limiting be added in any future versions? I see it in >>> the 4.0 feature requests, but not in any changelogs. >> >> I think the hope is that we can use the generic Linux mechanisms for >> this rather than having to add anything Xen-specific. >> >> J > _______________________________________________ Xen-devel mailing list Xen-devel@xxxxxxxxxxxxxxxxxxx <mailto:Xen-devel@xxxxxxxxxxxxxxxxxxx> http://lists.xensource.com/xen-devel -- */ Founder | Director | VP Research Orion Virtualisation Solutions/* | www.orionvm.com.au <http://www.orionvm.com.au/> | Phone: 1300 56 99 52 | Mobile: 0428 754 846 _______________________________________________ Xen-devel mailing list Xen-devel@xxxxxxxxxxxxxxxxxxx http://lists.xensource.com/xen-devel
|
Lists.xenproject.org is hosted with RackSpace, monitoring our |