Hi all, These are new releases of dm-ioband and blkio-cgroup. The major change of these releases is that dm-ioband handles sync/async IO requests separately and it solves write-startve-read issue pointed out by Vivek. Thank you Vivek for your pointing out the issue. Subject: ioband: Writer starves reader even without competitors http://lkml.org/lkml/2009/9/15/478 Here is the test result. I did the same test as Vivek did, it reads 32MB files four times during a bufferd writer is running. dm-ioband v1.14.0 file # 1, plain reading it took: 0.73 seconds file # 2, plain reading it took: 0.78 seconds file # 3, plain reading it took: 0.78 seconds file # 4, plain reading it took: 0.94 seconds dm-ioband v1.13.0 file # 1, plain reading it took: 14.27 seconds file # 2, plain reading it took: 22.17 seconds file # 3, plain reading it took: 12.31 seconds file # 4, plain reading it took: 17.69 seconds The summary of the changes are below: dm-ioband v1.14.0 - Handle sync/async IO requests separetely, it solves write-starve-read issue. - Allow it to compile against 2.6.32-rc1. blkio-cgroup v13 - Fix style issues caught by checkpatch.pl. - Allow it to compile against 2.6.32-rc1. The list of the patches: [PATCH 1/9] I/O bandwidth controller and BIO tracking [PATCH 2/9] dm-ioband-1.14.0: All-in-one patch [PATCH 3/9] blkio-cgroup-v13: The new page_cgroup framework [PATCH 4/9] blkio-cgroup-v13: Refactoring io-context initialization [PATCH 5/9] blkio-cgroup-v13: The body of blkio-cgroup [PATCH 6/9] blkio-cgroup-v13: Page tracking hooks [PATCH 7/9] blkio-cgroup-v13: The document of blkio-cgroup [PATCH 8/9] blkio-cgroup-v13: Add a cgroup support to dm-ioband [PATCH 9/9] blkio-cgroup-v13: The document of a cgroup support for dm-ioband About dm-ioband dm-ioband is an I/O bandwidth controller implemented as a device-mapper driver and can control bandwidth on per partition, per user, per process, per virtual machine (such as KVM or Xen) basis. About blkio-cgruop blkio-cgroup is a block I/O tracking mechanism implemented on the cgroup memory subsystem. Using this feature the owners of any type of I/O can be determined. This allows dm-ioband to control block I/O bandwidth even when it is accepting delayed write requests. dm-ioband can find the cgroup of each request. It is also for possible that others working on I/O bandwidth throttling to use this functionality to control asynchronous I/O with a little enhancement. Please visit our website, the patches and more information are available. Linux Block I/O Bandwidth Control Project http://sourceforge.net/apps/trac/ioband/ I'd like to get some feedbacks from the list. Any comments are appreciated. Thanks, Ryo Tsuruta -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
