similar to: mount point is not unique among all nodes

Now the discussion moves to kernel space, I move the email from ocfs2-tools-devel to ocfs2-devel. The original discussion can be found from http://oss.oracle.com/pipermail/ocfs2-tools-devel/2009-June/001891.html Joel Becker Wrote: > On Sat, Jun 27, 2009 at 03:46:04AM +0800, Coly Li wrote: >> Joel Becker Wrote: >>> On Sat, Jun 27, 2009 at 03:00:05AM +0800, Coly Li wrote: >>

All, Please review the patches backported to 1.4 from mainline. Sunil

Hello, I've hit a bug in OCFS2 delete code which results in inodes being left on disk without any links to them. The workload triggering this creates directories on one node and deletes them on another node in the cluster. The inode is not deleted because both nodes bail out from ocfs2_delete_inode() with: Skipping delete of 100405 because it is in use on other nodes The scenario which I

Found 15 patches (out of 162) that appeared relevant to ocfs2 1.4. Please review. Sunil

Using fsdlm/ocfs2_controld.cman, I've rerun the test I've been having problems with on 2.6.30-rc1. After running for several minutes in the same directory on three nodes, the test hangs, and I collect the following information: bull-01 ------- 3053 S< [ocfs2dc] ocfs2_downconvert_thread 3054 S< [dlm_astd] dlm_astd 3055 S< [dlm_scand]

Currently f_fsid of struct kstatfs returned from ocfs2_statfs() is undefined (vfs layer fills 0 as default). Since in some conditions, f_fsid value might be used as (f_fsid, ino) pair to uniquely identify a file, ocfs2 should return a unique defined f_fsid value from ocfs2_statfs(). Because uuid_str is identified no mater on big or litlle endian machine, it's also endian consistent to use

Resending after implementing review comments.

On unlink, all nodes check for the dentry in dcache and if present they mark the node as unlinked. The last node that purges the inode will clean it from orphan directory. When there is a memory pressure, a dentry may not be around and hence the inode is not marked as deleted and this will lead the file to be in the orphan directory till the slot is re-used during next mount. This patch initiates

Implemented the suggestions. Also, moved the ocfs2_orphan_scan_init() to the end of ocfs2_fill_super() instead of calling it earlier in ocfs2_initialize_super(). Sunil

The Lock Value Block (LVB) of a DLM lock can be lost when nodes die and the DLM cannot reconstruct its state. Clients of the DLM need to know this. ocfs2's internal DLM, o2dlm, explicitly zeroes out the LVB when it loses track of the state. This is not a standard behavior, but ocfs2 has always relied on it. Thus, an o2dlm LVB is always "valid". ocfs2 now supports both o2dlm and

Hi Joel/Srini, Here are 2 patches for orphan lock in local mode. patch 1: In local mode, we don't need lvb, so don't init it. patch 2: In local mode, orphan lock and unlock don't need to go to dlm part. Regards, Tao

I observer interesting dentry number during the dlm stress testing. The ocfs2 cluster stack is user space "pcmk". creating a empty directory, ls -al: total 8 drwxr-xr-x 2 root root 4096 2009-05-08 15:29 . drwxr-xr-x 4 root root 4096 2009-05-08 15:29 .. run script make_panic on single node, for 10 seconds, ls -al: total 12 drwxr-xr-x 2 root root 4096 2009-05-08 15:33 . drwxr-xr-x

Resending after implementing review comments.

Currently f_fsid of struct kstatfs returned from ocfs2_statfs() is undefined (at least it should be filled with 0). Since in some conditions, f_fsid value might be used as (f_fsid, ino) pare to uniquely identify a file, ocfs2 should return a defined unique f_fsid value from ocfs2_statfs(). This patch uses uuid_hash as a unique ID to initiate f_fsid value, the 32bits width is enough for ocfs2

Next two patches are backport of orphan scan timer patches to ocfs2-1.4

I'm running a bit late with the e-mail this time around, but I think that's ok since there really isn't any major new features here - the bulk of the Ocfs2 update is bug fixes, or cleanups. The same goes for configfs. The only two things that could be described as features would be: - Sunil has updated Ocfs2 to provide even more live cluster locking information via debugfs. - Joel

Resending after adding another patch to display delayed orphan scan statistics.

This patch adds jbd_lock_bh_state() and jbd_unlock_bh_state() around accessses to jh->b_committed_data. Fixes oss bugzilla#1131 http://oss.oracle.com/bugzilla/show_bug.cgi?id=1131 Signed-off-by: Sunil Mushran <sunil.mushran at oracle.com> --- fs/ocfs2/suballoc.c | 28 ++++++++++++++++++++++++---- 1 files changed, 24 insertions(+), 4 deletions(-) diff --git a/fs/ocfs2/suballoc.c

Patch 5 has been respun with the recommended change... remove the ex argument in ocfs2_orphan_scan_lock() and ocfs2_orphan_scan_unlock(). Sunil

In fs/ocfs2/dlmglue.c:ocfs2_cluster_lock(), after label 'out:' the code is: 1373 if (wait && arg_flags & OCFS2_LOCK_NONBLOCK && 1374 mw.mw_mask & (OCFS2_LOCK_BUSY|OCFS2_LOCK_BLOCKED)) { 1375 wait = 0; 1376 if (lockres_remove_mask_waiter(lockres, &mw)) 1377 ret = -EAGAIN; 1378