On 06/10/2013 07:35 PM, Stephane Boisvert wrote: > Hi, > I wondering how safe it is to use rbd cache = truewith libvirt/qemu. > I did read the documentation and it says "When the OS sends a barrier > or a flush request, all dirty data is written to the OSDs. This means > that using write-back caching is just as safe as using a well-behaved > physical hard disk with a VM that properly sends flushes" > > but what does it really means.. "just as safe as physical hard drives" I > don't know how safe are physical hard drives.. Is it safe to enable that > setting on a prod environment ? If I don't enable that cache I'm just > getting really poor write performance on the VMs. I have a working prod-environment, and I'm using it with caches, it works absolutely fine. The trick is, that the OS can send two different write requests. One is: "write, and I don't care when you finish writing, here's your data". The other one is: "write, but I want to wait until all data is on disk (sync)". Ceph Cache behaves as it should, meaning the first type of write is written when there's time for it, and the other one is written immediately. > What are the implications of enabling the caching when using > libvirt/qemu/kvm ? you should really upgrade to a recent qemu-version, because the qemu / rbd layer had some bugs concerning caches and high write load. > > Thanks hope this helps Wolfgang > _______________________________________________ > ceph-users mailing list > ceph-users@xxxxxxxxxxxxxx > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com -- DI (FH) Wolfgang Hennerbichler Software Development Unit Advanced Computing Technologies RISC Software GmbH A company of the Johannes Kepler University Linz IT-Center Softwarepark 35 4232 Hagenberg Austria Phone: +43 7236 3343 245 Fax: +43 7236 3343 250 wolfgang.hennerbichler@xxxxxxxxxxxxxxxx http://www.risc-software.at _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |