Re: Re: [PATCH] ceph:Update the file time after mmap-write.

>On Thu, Jul 11, 2013 at 9:17 AM, majianpeng wrote: >> Although, mmap-write of ceph update the time of file using >> file_update_time.Because it don't mark the relative cap so the time >> can't save. > >I think cephfs' mmap IO support is still broken. mmap IO does not respect >ceph capabilities at

Re: 4x write amplification?

2013/7/10 Gregory Farnum : > On Tue, Jul 9, 2013 at 7:08 PM, Li Wang wrote: >> Hi, >> We did a simple throughput test on Ceph with 2 OSD nodes configured >> with one replica policy. For each OSD node, the throughput measured by 'dd' >> run locally is 117MB/s. Therefore, in theory, the two OSDs c

[PATCH] ceph: Don't get layout from mds for regular file which has been written.

For regular file which has been written, it can't change layout. So it can get layout from local rather than mds. Signed-off-by: Jianpeng Ma --- fs/ceph/ioctl.c | 10 -- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c index e0b4ef3..2b74879

Re: Re: [PATCH] ceph:Update the file time after mmap-write.

On Thu, Jul 11, 2013 at 3:53 PM, majianpeng wrote: >>On Thu, Jul 11, 2013 at 9:17 AM, majianpeng wrote: >>> Although, mmap-write of ceph update the time of file using >>> file_update_time.Because it don't mark the relative cap so the time >>> can't save. >> >>I think cephfs' mmap IO support is st

Re: Set object mtime

Hi, > Okay thanks! A call in the C API would be handy. I was wanting to look > at creating a tool to sync RADOS between clusters. Is that anything > that's in the development plan already? I was just thinking of the same thing. Basically I want rsync between pools, mostly to allow changing the n

Re: wip-libcephfs-rebased-v65 ready for pull

Hi, I've rebased this on master, and cherry-picked the fixes from Ilya on top. The exception is the xattr changes, which turned out to be an MDS bug turned by up the unit test. Is there a particular need for this to be in dumpling? I'm inclined to delay a release to avoid potentially destabl

Re: [PATCH] ceph: Don't do set_layout for file which has been written.

On Thu, 11 Jul 2013, majianpeng wrote: > Cephfs don't change the layout which has been written.So add this check > before do. This could conceivaably change in the future, and I'm not sure there is harm is the MDS round trip to get the error code here...? sage > > Signed-off-by: Jianpeng Ma >

Re: wip-libcephfs rebased and pulled up to v.65

Please check out the current wip=libcephfs branch and let me know how it looks/works for you guys. I cleaned up your patches a bit and fixed teh root cause of the xattr issue you were seeing. Thanks! sage On Thu, 11 Jul 2013, Ilya Storozhilov wrote: > Hi Adam (CC: Sage, Matt and David), > >

Re: wip-libcephfs-rebased-v65 ready for pull

Hi Sage, It's up to you folks. If the next window works better, let's just enter it early. Matt - "Sage Weil" wrote: > Hi, > > I've rebased this on master, and cherry-picked the fixes from Ilya on > top. > The exception is the xattr changes, which turned out to be an MDS bug > > turn

Re: Using valgrind_osd option

On Wed, 10 Jul 2013, Ranjan Sarpangala Venkatesh wrote: > Hi Mark, > > I upgraded valgrind to 3.8.1 and also used > --soname-synonyms=somalloc=*tcmalloc* option. However, no massif.out > file was created. > > Kindly let me know if i am missing something. I haven't run into this. I usually run w

Re: Using valgrind_osd option

Hrm, I think I answered too quickly. I use that directly with valgrind launching the daemon, not through vstart. Not sure if it passes it through properly. Maybe try Sage's suggestion or see if valgrind works when you launch the daemon directly? Mark On 07/11/2013 01:21 AM, Ranjan Sarpanga

Re: Re: [PATCH] ceph: Don't do set_layout for file which has been written.

>On Thu, 11 Jul 2013, majianpeng wrote: >> Cephfs don't change the layout which has been written.So add this check >> before do. > >This could conceivaably change in the future, and I'm not sure there is >harm is the MDS round trip to get the error code here...? > Sorry, i can't understand what's

Re: Re: [PATCH] ceph:Update the file time after mmap-write.

>On Thu, Jul 11, 2013 at 3:53 PM, majianpeng wrote: >>>On Thu, Jul 11, 2013 at 9:17 AM, majianpeng wrote: Although, mmap-write of ceph update the time of file using file_update_time.Because it don't mark the relative cap so the time can't save. >>> >>>I think cephfs' mmap IO suppor

Re: Re: [PATCH] ceph:Update the file time after mmap-write.

On Thu, Jul 11, 2013 at 2:03 AM, Yan, Zheng wrote: > On Thu, Jul 11, 2013 at 3:53 PM, majianpeng wrote: >>>On Thu, Jul 11, 2013 at 9:17 AM, majianpeng wrote: Although, mmap-write of ceph update the time of file using file_update_time.Because it don't mark the relative cap so the time >

Re: Re: [PATCH] ceph:Update the file time after mmap-write.

On Fri, Jul 12, 2013 at 10:46 AM, Gregory Farnum wrote: > On Thu, Jul 11, 2013 at 2:03 AM, Yan, Zheng wrote: >> On Thu, Jul 11, 2013 at 3:53 PM, majianpeng wrote: On Thu, Jul 11, 2013 at 9:17 AM, majianpeng wrote: > Although, mmap-write of ceph update the time of file using > file_u

HA: wip-libcephfs rebased and pulled up to v.65

Hi Sage, Adam, Matt and David, this branch is looking good at the first blush. We will check out how it works and report to you till the EOD. Regards, Ilya От: Sage Weil [s...@inktank.com] Отправлено: 11 июля 2013 г. 23:01 To: Ilya Storozhilov Cc: Adam C.

Re: Using valgrind_osd option

Hi Sage,Mike I reverted back to valgrind version 3.7.0, did what Sage suggested and it worked. However, with valgrind 3.8.1 no massif.out file was created even while running a build with no tcmalloc. Not sure why this is happening. Thanks Sage and Mark for your help. With regards Ranjan On Thu,