Date: 2010-11-19
OS: SLES
OS Version: 10 SP2
Veritas Storage Foundation Scalable File Server (now called Symantec FileStore)
5.5 SP1RP2HF2 Patch Upgrade README

CONTENTS
I.   Overview
II.  Fixes in the new patch
III. Known issues
IV.  Upgrade procedure

================================================
I. Overview
================================================
The Veritas Storage Foundation Scalable File Server (now called Symantec FileStore) provides a scalable clustered storage solution. This document provides release information for the patch.

================================================
II. Fixes in the new patch
================================================
Etrack Incidents: 2081313, 2081320, 2081324, 2085791, 2086029, 2087178, 2087240, 2087327, 2087560, 2092454, 2097392, 2105097, 2113769, 2125400, 2125612, 2125669, 2125857, 2126923, 2128190, 2128204, 2134790, 2134848, 2135925, 2138526, 2147580, 2148653, 2148656, 2149023, 2153045, 2158046, 2158137, 2161564, 2168975, 2168976, 2168984, 2168986, 2168988, 2168995, 2169001, 2169004, 2169022, 2169217, 2169224, 2169226, 2169227, 2169232, 2169241, 2171178, 2174511, 2175860, 2176812, 2177015, 2180447, 2180449, 2191998

Errors/Problems Fixed: Listed below.
2081313 Create snapshot appears to work successfully when the file system utilization rate at 100%, but no snapshot was actually created.
2081320, 2169004	A snapshot is created even though the number exceeded the maximum snapshot limit.
2081324 Auto snapshot can not be removed automatically even when the file system is full.
2085791 Data of snapshot schedule and tier schedule can not be set successfully with several discrete time points.
2086029 Data is different after replication.
2087178 Ensure integrity of replicated data on the remote site.
2087240, 2087560 The Replication > job resync and trigger commands behaves unexpectedly after the job is aborted. A new state, Enable-after-Aborted, should be added.
2087327 WAN trouble causes a Job Failed error after 30-40 minutes and the Replication > job disable command is the only command to restore this state.
2092454	In a homedir share, the Quota Limit disappears when a set quota value is configured for a new user with access from a Windows client. 
2097392 Replication fails to copy a multilayer directory (50 or more layers) from source to destination.
2105097	The destroyall command hangs and causes other commands to hang.
2113769	Four jobs can not be enabled successfully at the same time.
2125400	_nlm_ offline leads to partial status for a replication service.
2125612	Auto-snapshot can not be created according to the schedule.
2125669	Auto-snapshot can not be created when all cables are offline. This causes the Console IP to go offline.
2125857	Attribute access time of files in the source can not be synchronized to the destination.
2126923 After deleting the secondary tier of a file system, a "Command exited due to unknown error." message appears.
2128190	Client fails to refresh NFS share information.
2128204 The physical IPs and Virtual IPs can be in different subnet.
2134790	If the softquota is larger than the hardquota, a success message appears, but the actual operation fails.
2134848	When the soft quota option of the CIFS> homedir share command is set for the first time, the limit function fails.
2135925	Spelling error with patch install message.
2138526	Job can not be enabled because the checkpoint can not be mounted.
2147580	When the number of VIPs is more than forty-five , the client can not get access to the file system after a CIFS split.
2148653	Ip rule not created for bond0.10 device.
2148656	RPC tasks consume a large amount of memory.
2149023	The debuginfo command does not capture FTP logs in the /var/log directory.
2153045	The Storage> quota cifshomedir showdefault command displays values in the wrong columns.
2158046	When a CIFS share is set to guest and the share is split, accessing the share requires a password.
2158137	Accept the SNAPSURE EMC celerra environment variable.
2161564	CIFS service did not failover properly. All [_cfname$] parameters are gone after all pubeth are pulled off.
2168975 After setting the quota for a CIFS local user, the correct value is not displayed.
2168976	CIFS user quota input parameter has no length limit.
2168984 When executing a Replication> job disable command, command status includes extraneous scripts information.
2168986	Client cannot access the CIFS homedir share after the network link is down, then reconnected.
2168988	When the file system is full, files can not be deleted and  the file system can not be destroyed.
2168995 After applying Sept 14th patch, master node fails. 
2169001 When the Storage> create pool commands fails, a "Can't find disk" error message appears.
2169022	The CIFS > quota homedir softlimit command fails.
2169217	After deleting a CIFS normal share, you can still read and write the share.
2169224 Can not update disk information when the /var directory is full.
2169226	While CIFS share is set to the guest mode, the client computer accessing a CIFS share file requests a user name and password after the CIFS share is split.
2169227	For more than 45 VIPs, the client can not access the share after executing the split command.
2169232	The Space Used information in the CIFS> homedir group quota command is not updated.
2169241	After a file system is created, mount fails on some nodes.
2171178	NDMP restore fails for single files larger than 2 GB.
2174511 When you enable a user or group quota limit on a 12T file system, the command hangs for an hour and you can not access the file system.
2175860	NLM postonline interprets regular files in the /var/lib/nfs/sm directory as host names.
2176812	The Auto snapshot max_limit should never be exceeded.
2177015	Change the output for the NDMP> config_get_server_info command and the NDMP> config_get_host_info command to include detailed OS name and version information.
2180447 Checkpoint leftover on source side.
2180449	Replication jobs that create a large number of files causes an "Out of Memory" error on the source side.
2191998 Replication core files appeared if the absolute path of files are very long.

================================================
III. Known issues
================================================
2176889 Files copied from CIFS share are opened as directories.

Workaround: 
To copy files through a CIFS share when you do replication, add the option "no_full_acl" for the CIFS share. For example, CIFS> share add fs1 share1 no_full_acl, guest, rw...

2153118 The CIFS> map user command does not work as expected.

Workaround:
The domain you specify for CIFS user name mapping must be the netbios domain name (instead of the Active Directory DNS domain name) for the user. For example, a netbios domain name might be listed as SYMANTECDOMAIN instead of SYMANTECDOMAIN.COM (without the .COM extension).

To determine the netbios domain name, login to your Active Directory Server and type the following in a command window:

set | findstr DOMAIN

The results will include:

USERDOMAIN netbios_domain_name
USERDNSDOMAIN Active_Directory_DNS_domain_name

Use the value of the USERDOMAIN (the netbios domain name) when you map user names.

================================================
IV. Upgrade Procedure
================================================
NOTE: After you have installed, uninstalled, or synchronized a new SFS patch into your cluster, the list of available commands may change. Please re-login to the CLI to access the updated features. Since this is a phased upgrade process, please bring down the services manually and then start with the patch installation. After you apply this patch, you can not uninstall it.

Install the patch:
1. Login as master:
   su - master
2. Start the patch install:
   upgrade patch install <URL>