5.6.1 HP X9000 Series Release Notes (TA768-96053, September 2011)

Fixes
The following fixes were made in this release:
File systems
The ibrix_fsck command needed an option to focus consistency checking on specific segments.
A modification to the lost+found directory could cause corruption, with the error
ext3_dx_find_entry: bad entry in directory.
A race condition caused an error when ibrix_fsck checked the lost+found directory.
Running ibrix_fsck in phase 2 caused the file system to be read-only for X9000 clients.
Under certain conditions, mounting a file system could cause a server to terminate unexpectedly.
The mtime of a file changed when the file was migrated, rebalanced, or evacuated to another
segment.
A segment did not activate for a file system, causing an exception in JmsResource.
CIFS
Applications connecting to X9000 CIFS shares experienced file read delays when oplocks were
enabled.
A CIFS share could not be deleted.
Performance improvements were needed for CIFS and Windows X9000 clients.
A segmentation fault occurred after the CIFS service was restarted.
If the guest account did not have a password, it was disabled when the CIFS service was restarted.
A race condition caused intermittent CIFS outages.
A condition in the CIFS implementation could cause an lsassd failure and segmentation fault.
The CIFS implementation did not support command chaining.
A condition in the CIFS implementation could cause CIFS disconnects and lsassd failures.
When extremely large numbers of files were opened, CIFS users lost access to the CIFS share.
A condition in the CIFS implementation caused a backup application to fail to write a file.
The CIFS implementation failed with lwiod segfault in IbfsSyncLockState().
When Cygwin was used to create a hard link on a CIFS share, the CIFS implementation did not
report that hard links are not supported.
If a CIFS client had a file open on an X9000 CIFS share and the client then lost network access
without closing the file gracefully, the file could not be opened by other CIFS clients until a timeout
period expired. This condition has been resolved by changing the TCP timeouts used by the CIFS
implementation. If you need to further adjust the timeout periods, locate the following parameters
in the Likewise registry:
"TcpKeepaliveCnt"=dword:00000009
"TcpKeepaliveIdle"=dword:00000618
"TcpKeepaliveIntvl"=dword:0000001e
Fixes 5