Re: [PATCH] ceph: Add FScache support

2013-07-09 Thread David Howells
Milosz Tanski wrote: > It looks like both the cifs and NFS code do not bother with any > locking around cifs_fscache_set_inode_cookie. Is there no concern over > multiple open() calls racing to create the cookie in those > filesystems? Yeah... That's probably wrong. AFS obviates the need for s

Re: [PATCH] ceph: Add FScache support

2013-07-09 Thread Myklebust, Trond
On Tue, 2013-07-09 at 11:33 +0100, David Howells wrote: > Milosz Tanski wrote: > > > It looks like both the cifs and NFS code do not bother with any > > locking around cifs_fscache_set_inode_cookie. Is there no concern over > > multiple open() calls racing to create the cookie in those > > filesy

Re: [PATCH] ceph: Add FScache support

2013-07-09 Thread Milosz Tanski
On Tue, Jul 9, 2013 at 6:33 AM, David Howells wrote: > Milosz Tanski wrote: > >> It looks like both the cifs and NFS code do not bother with any >> locking around cifs_fscache_set_inode_cookie. Is there no concern over >> multiple open() calls racing to create the cookie in those >> filesystems?

Re: [PATCH] ceph: Add FScache support

2013-07-09 Thread Milosz Tanski
David, On a somewhat related node. The header definition of fscache_maybe_release_page doesn't seam quite correct. The comment states it should return true if the page can be freed ... yet when there's not a cookie or PG_fscache is not set it returns false. This doesn't seam right, in fact the NFS

Re: [PATCH] ceph: Add FScache support

2013-07-09 Thread Milosz Tanski
David, I have another one for you... I though maybe I caused this somehow so I spent like 8 hours tracking. But now I'm not so sure. I got this error when running this "stress test" root@pbnode-58019a39:/mnt/cluster/petabucket/prod# find -type f | xargs -P 16 cat | pv > /dev/null ^C83GB 0:00:52

[PATCH] Python Ioctx.get_xattrs does not take xattr_name

2013-07-09 Thread Johannes Erdfelt
Fixes bug 5528 Since it returns all xattrs, it does not take an xattr_name. Also, add unit tests for the Object xattrs methods. Signed-off-by: Johannes Erdfelt --- src/pybind/rados.py | 4 ++-- src/test/pybind/test_rados.py | 12 2 files changed, 14 insertions(+), 2 dele

assertion failure in update_from_paxos

2013-07-09 Thread Noah Watkins
I'm getting the following failure when running a vstart instance with 1 of each daemon. -- 0> 2013-07-09 08:30:43.213345 7fdc289e97c0 -1 mon/OSDMonitor.cc: In function 'virtual void OSDMonitor::update_from_paxos(bool*)' thread 7fdc289e97c0 time 2013-07-09 08:30:43.207686 mon/OSDMonitor.cc: 129: F

Re: [PATCH] ceph: Add FScache support

2013-07-09 Thread Milosz Tanski
It turns out that I did this to myself with the launder_page callback that implemented. Originally I would be able to solve the occasional page left over with PG_fscache mark. - Milosz On Tue, Jul 9, 2013 at 10:07 AM, Milosz Tanski wrote: > David, > > I have another one for you... I though maybe

v0.66 released

2013-07-09 Thread Sage Weil
Our last development release before dumpling is here! The main improvements here are with monitor performance and OSD pg log rewrites to speed up peeering. In other news, the dumpling feature freeze is upon us. The next month we will be focusing entirely on stabilization and testing. There w

[GIT PULL] Ceph updates for 3.11

2013-07-09 Thread Sage Weil
Hi Linus, Please pull the following Ceph updates from git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client.git for-linus There is some follow-on RBD cleanup after the last window's code drop, a series from Yan fixing multi-mds behavior in cephfs, and then a sprinkling of bug fixes

Re: assertion failure in update_from_paxos

2013-07-09 Thread Joao Eduardo Luis
On 07/09/2013 04:37 PM, Noah Watkins wrote: I'm getting the following failure when running a vstart instance with 1 of each daemon. I can confirm this happens, as it just happened to me as well. My guess is that this is something Sage may have fixed last night, but will have to check. -Jo

Re: assertion failure in update_from_paxos

2013-07-09 Thread Noah Watkins
It appears to be resolved in master now. On Tue, Jul 9, 2013 at 12:43 PM, Joao Eduardo Luis wrote: > On 07/09/2013 04:37 PM, Noah Watkins wrote: >> >> I'm getting the following failure when running a vstart instance with >> 1 of each daemon. > > > I can confirm this happens, as it just happened t

Re: assertion failure in update_from_paxos

2013-07-09 Thread Joao Eduardo Luis
On 07/09/2013 08:45 PM, Noah Watkins wrote: It appears to be resolved in master now. Yep! -Joao On Tue, Jul 9, 2013 at 12:43 PM, Joao Eduardo Luis wrote: On 07/09/2013 04:37 PM, Noah Watkins wrote: I'm getting the following failure when running a vstart instance with 1 of each daemon.

Using valgrind_osd option

2013-07-09 Thread Ranjan Sarpangala Venkatesh
Hi, I am trying to collect memory of the OSD daemon with the following ./vstart.sh -n -x -l --valgrind_osd 'massif' However, when I print massif.out file there is nothing cat massif.out.15391 desc: (none) cmd: ./ceph-osd -i 0 -c ceph.conf -f time_unit: i #--- snapshot=0 #--- ti

Re: Using valgrind_osd option

2013-07-09 Thread Mark Nelson
Hi, I ran into this problem a while back and it was because I was trying to use a version of valgrind that didn't support tcmalloc (which Ceph was compiled to use). Try upgrading valgrind to 3.8.0 or newer. You'll also need to pass something like: --soname-synonyms=somalloc=*tcmalloc* Mar

4x write amplification?

2013-07-09 Thread Li Wang
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 could provide 200+MB/s throughput. However, using 'iozone' from clients we only get a