Date: 2012-04-25 OS: SLES OS Version: 11 SP1 Symantec FileStore 5.7 P2 Patch Upgrade README CONTENTS I. Overview II. Upgrade procedure III. Fixes in the new patch IV. Known issues V. New Features VI. Deprecated Feature PATCH ID : N/A PATCH NAME : sfs-sles11_x86_64-5.7P2-patches.tar.gz BASE PACKAGE NAME : Symantec FileStore BASE PACKAGE VERSION : 5.7 OBSELETE PATCHES : N/A SUPERCEDED PATCHES : N/A INCOMPATIBLE PATCHES : N/A SUPPORTED OS : SLES SUPPORTED OS VERSION : SLES 11 SP1 CREATION DATE : 2012-04-25 CATEGORY : enhancement, performance issue REBOOT REQUIRED : Yes SUPPORTS ROLLBACK : NO I. OVERVIEW: ------------ Symantec FileStore provides a scalable clustered storage solution. This document provides release information for the patch. Notes: 1. Patch upgrade will cause service failover from one node to another which will take about 15 seconds. This will not impact NFS client mounts and replication jobs, but CIFS client mounts will need to connect again, and backup jobs need to be resumed as well after an upgrade. 2. For the above reason, it is recommended to stop CIFS, backup, and EV I/O loads during an upgrade. 3. The 5.7 P2 patch can only be applied to the 5.7 or 5.7 P1 release. 4. Patch upgrade cannot be rollbacked once applied. II. UPGRADE PROCEDURE: ---------------------- To install the patch: 1. Copy SFSFS-patch-5.7P2_rc9_2012_04_25.tar.gz to the FileStore master node at (/home/support/patches/) using the support login. 2. If fencing is enabled, turn off and remove I/O fencing before the upgrade: filestore> storage> fencing off filestore> storage> fencing destroy 3. Install the FileStore patch using the master login (CLISH): filestore>upgrade patch install scp://support@localhost:/home/support/patches/SFSFS-patch-5.7P2_rc9_2012_04_25.tar.gz The patch is installed on the slave nodes first followed by a system reboot. The master node is installed last followed by a system reboot. 4. If fencing was enabled, enable it again. 5. Monitor and check the status of the FileStore cluster: To check that the installation is complete, login to the master node again (switch to another node), and check that all the nodes are in the RUNNING state: filestore> cluster show ------------ ----- ----------- -------- -------- -------- -------- filestore_01 RUNNING 0.98 0.01 0.00 0.01 0.00 filestore_02 RUNNING 0.97 0.01 0.00 0.01 0.00 filestore_01# gabconfig -a GAB Port Memberships ================================= Port a gen 187002 membership 01 Port b gen 187004 membership 01 Port f gen 18700e membership 01 Port h gen 187006 membership 01 Port u gen 18700c membership 01 Port v gen 187008 membership 01 Port w gen 18700a membership 01 Note: Please refer to the Symantec FileStore 5.7 Command-Line Administrator's Guide for information on the "support" user login and more information about the patch installation. III. FIXES IN THE NEW PATCH: ---------------------------- Etrack Incidents: 2698792, 2700679, 2706400, 2703052, 2715479, 2674475, 2644052, 2737157, 2705765, 2691194, 2741277, 2641131, 2707944, 2250569, 2635189, 2646601, 2609825, 2625248, 2640734, 2589988, 2710222, 2679385, 2698087, 2698019, 2684830, 2697885, 2684894, 2636506, 2615754, 2663303, 2383574, 2645760, 2522726, 2644807, 2657517, 2668615, 2724581, 2422305, 2668867, 2620457, 2715232, 2697465, 2705748, 2705740, 2682386, 2623551, 2618933, 2723024, 2717357, 2703188, 2644785, 2630158, 2697934, 2631136, 2645579, 2619813, 2655304, 2710879, 2583245, 2662850, 2697192, 2704863, 2736968, 2709768, 2634065, 2623121, 2685044, 2648788, 2661272, 2732163, 2204320, 2667775, 2639334, 2347003, 2660166, 2605884, 2606454, 2655342, 2416207, 2594441, 2648756, 2687975, 2653945, 2281709, 2657228, 2342588 Errors/Problems Fixed: 2698792 Event logs output on GUI was not sorted on 5.7P1 2700679 update_parameters.sh hangs indefinitely. 2706400 GUI>AntiVirus>LiveUpdate and Quarantine cannot be updated to current by a refresh. They should be updated on a timely basis. 2703052 GUI does not support new cifs share option, sharename@VIP, and other three issue related to GUI>share>cifs. 2715479 "logrotate: error: samba-winbind duplicate log" comes out periodically. Setting logrotate for samba should be fixed. 2674475 SNMP daemon goes into infinite loop filling /var in 5.7. SusE already has a fix for the same. We need to include the latest daemon. 2644052 Need to add release information in OS & FileStore tar ball. 2737157 Upgrade vxfs to VRTSvxfs-5.1.132.057-SP1RP2HFg_SLES11.x86_64.rpm 2705765 Need to improve the EVENT log rotation and cleanup. 2691194 [CIFS] Support AD idmap backend 2741277 Upgrade vcs to VRTSvcs-5.1.132.002-SP1RP2HF2_SLES11.i686.rpmssss. 2641130, 2641131 Ability to swap IP address of a physical NIC. 2707962, 2707944 Create a file system that is 11M in size. CLISH displays the following error message: "SFS fs ERROR V-288-902 Size specified for filesystem fs1 is less than 10MB". 2252612, 2250569 Rebooted FileStore cluster. However it came up and the master account was unavailable. 2639049, 2635189 Existing segregated share error message seen when attempting to reuse a share name. 2647155, 2646601 FILESTORE:DEDUP: debuginfo should not collect files inside mounted checkpoint. 2089620, 2609825 FILESTORE:SAV: "Reboot all" causes SAV panic on cluster node while no load was running. 2625932, 2625248 FILESTORE[STORAGE]: File system upgrade failed from v7 to v8. 2641042, 2640734 FILESTORE[STORAGE]: cvm sg did not switch to slave node. 2590239, 2589988 Fastresync became disabled after shrinking a striped-mirrored file system. 2710229, 2710222 FileStore: NFS/FTP/Backup Some of the commands are slow in responding. 2679382, 2679385 FileStore: NFS: Add manage-gids options to mountd, workaround for 16-group limit. 2698081, 2698087 FileStore: NFS: SLES11 NLM nat connections cannot be established through devices other than pubeth0 (default). 2407637, 2698019 FileStore: NFS: offline of NFS group takes a long time when there are a lot of shares (2 seconds per share). 2610201, 2684830 FileStore: iSCSI: column mis-alignment in output of iscsi target list. 2697881, 2697885 FileStore:NFS: Make NLM one-way udp nat for mac as default. 2639650, 2684894 Filestore::ISCSI::target attr initial_login_retry_max is getting set to -1 if we try to set it with a large value. 2637092, 2636506 Filestore:Storage:sfsdg not imported on the master node after a reboot. 2353265, 2615754 File system is not allowed to grow if usage is above 95%. 2663460, 2663303 If dedup or dryrun is started from CLISH login console, dedup/dryrun exits on console logout/exit. 2384451, 2383574 In system, unable to modify vxtune volpagemod_max_memsz to 1000000. 2620373, 2645760 Instructions/Utility needed for adding a new NIC to N8300 and moving console to dedicated management network on the new NIC. 2645983, 2522726 Maximize rcqsize during "fs create" for optimal dedup performance. 2644804, 2644807 Resolve issue of "NLM getting used up by events" by dropping event table and picking events from event log file. 2657516, 2657517 Rollback create failed when there is only one disk in the pool. 2668394, 2668615 When GUI database becomes offlined, the run_sfs_agentlet_worker.pl script still keeps running. 2725674, 2724581 Unable to add ftp local user. 2425044, 2422305 [CIFS] Need to handle netbios name change. 2669850, 2668867 [CIFS] Postonline and postoffline CIFS server and share at the same time. 2620455, 2620457 [CIFS] usability improvement on setting data migration option. 2697513, 2715232 FileStore 5.7P1 Web GUI requires SMTP username which is an optional attribute via CLISH. 2699303, 2697465 Some fields are missing in XML O/P when running the "storage dedup status FS_61" command. 2705772, 2705748 Provide a drop location for drivers with every installation. 2705797, 2705740 Provide silent installation facility for FileStore bits. 2682385, 2682386 SFS5.7P1: the 'Storage Summary' on GUI main page shows Free incorrectly as -991G. In certain cases, GUI shows pool over-allocation when there is none. 2623548, 2623551 /var/log/sfsfs_event.log does not have antivirus scan job detecting virus info on node_01, but it does have detecting virus information on node_02. 2099996, 2618933 Antivirus "Auto protect" does not detect virus when renaming virus-file to "11". "Auto protect" should detect virus in any filename. 2723020, 2723024 GUI cannot handle snapshot schedule name involving a "-" hyphen, but CLI can handle it. 2717355, 2717357 GUI, "FileSystem" does not show "Protocol(CIFS, NFS)" column properly in case of a share add for a sub-directory like /vx/NS1/dir1. 2608134, 2703188 GUI>File Systems>File Systems, click-FSname shows File System Details involving Tier Summary. Secondary Tier information is not updated by refresh button or `CLI>support gui db rescan` command. 2642218, 2644785 `cli>replication job pause` does not work during `replication job resync`. `replication job status` shows "paused" but `rsync` is still running by `ps` command. 2630156, 2630158 User wants to suppress TCPConnTrack connection new/close event logs because user has lots of messages (200, 000 events every day) 2696937, 2697934 Default number of nfsd is too large; primary node(ConIP node) becomes 100% busy very quickly. Default should be a small value like 8. 2631127, 2631136 Event message loss when management console switch over occurs. 2593659, 2645579 non_scan.tar.gz built from a certain simple/small text file causes "extraction error by Decomposer". Other files seem to get scanned. 2618306, 2619813 To exclude Samba from supportconfig. 2638369, 2655304 /var file system usage increase by the samba logs. 2710873, 2710879 [FileStore][NFS]: nodename argument missing in man pages for "option modify nfsd". 2583233, 2583245 [NETWORK] Ability to sort output by different columns. 2662848, 2662850 [Replication] Make replication rsync time unlimited. 2703580, 2697192 The cluster created lots of little files in the tmp directory when running automation. 2705780, 2704863 auto_snapshot cannot create successfully after the schedule of snapshot is set. 2715283, 2736968 The files of the secondary tier are moved incompletely on the primary tier after relocation is executed. 2708766, 2709768 When NDMP restored many files at one time, it will not restore successfully. 2634587, 2634065 Cannot access a homedir share if the slave node is powered off. 2623522, 2623121 Delete the VIP immediately after adding successfully, which reports deletion successfully, but the VIP still exists. 2650396, 2685044 HTTP MaxSpareThreads default is 75. But set MaxSpareThreads must be in the range of 1-64. 2649439, 2648788 If the network interface goes down in the process of providing service, the previous node will not release the smbd process at once. 2659364, 2661272 Cannot offline the FTP shared file system after stopping the FTP service. 2732159, 2732163 Reboot the slave node when doing NFS operations on both nodes. The master node was interrupted for about 10 minutes, and the nfsd of the master was killed after the slave rebooted successfully. 2217904, 2204320 Scan virus for a file system with a large number of files. The process rtvscand occupies a lot of CPU. 2668539, 2667775 The dns configure file on the master node was different on the slave node after the cluster installed successfully. 2634576, 2639334 The function of homedirfs default quota does not take effect. 2348938, 2347003 The tomcat service is automatically offlined after a few minutes. 2660508, 2660166 Configured the quota of users, which printed the wrong first character of the user's name. 2595596, 2605884 When enabling the cifshomedir quota service, the command hangs for half an hour (reproduce kswapd issue). 2351651, 2606454 Job failed after about 220 cycles without data changes. 2655903, 2655342 Job one is triggered after adding some files and deleting the else files on the file system of the source cluster. Job one failed. 2414709, 2416207 Allow change cluster name without a re-install. 2594881, 2594441 cifs share show error in ctdb mode. 2634579, 2648756 Modify winbind_monitor and dependency between CIFS and nlm file system. 2686764, 2687975 upgrade show takes >1min to complete. 2653944, 2653945 FileStore:Reports: Report Showevents shows junk Script messages in clish. 2281708, 2281709 [system] "system more disabled" does not work for "system clock regions America" 2657516, 2657228 Rollback create failed when there is only one disk in the pool 2336726, 2342588 log files filling file systems IV. KNOWN ISSUES: ----------------- Etrack: 2624703, 2631108 Symptom: Previously deduplicated files (or portions of the files) overwritten by identical content remain unshared. Description: If a previously deduplicated file (or a portion of it) is overwritten by identical content, the overwritten portions become unshared. Future deduplication is unable to deduplicate the file (or the overwritten portion). Workaround: This is a known limitation. For full-file overwrites, the workaround is to first delete the file that is to be overwritten. Etrack: 2645764 Symptom: Deduplication failed with "AddBlock failed. Error = 110". Description: AddBlock error 110 hits due to the low space in the file system. Low space might be caused by I/O on the file system or temporary space occupied by the deduplication. Workaround: Free space in the file system and re-run the deduplication. Etrack incident: 2733047 SYMPTOM: When starting a Storage> dedup or a Storage> dedup dryrun command on a file system, the following error is seen: storage dedup status Filesystem Saving Status Node Type Details -------------------------------------------------------------------------------- ------------- /vx/FS_61 42% FAILED NSTMP_01 MANUAL 2012/02/14 13:36:04 Deduplication mode error DESCRIPTION: From the error above, the user tried to do a dryrun on a file system that is already deduplicated. RESOLUTION: If the Storage> dedup job has been run on the file system at least once before, then the Storage> dryrun option would not work as is. Etrack incident: 2698672 SYMPTOM: When a VIP fails over to other nodes, CIFS clients may experience connection issues to FileStore. DESCRIPTION: When a VIP fails over to other nodes, the Windows client still caches the old entry of the DFS referral, causing the Windows client to try and connect to the old VIP. The windows client may experience connection denied issues in this case. RESOLUTION: Reboot the windows client to clear the DFS cache. Etrack: 2735401 SYMPTOM: Adding a mirror for the NFS Lock Management (NLM) module or Data Change Object (DCO) volumes can fail if there is only one storage pool with one disk on a FileStore cluster. DESCRIPTION: On FileStore, by default NLM and DCO volumes are mirrored for redundancy. When there is only one disk available, NLM or DCO mirror operations can fail. Moreover, since the configuration copies of Volume Manager cannot be distributed across multiple physical disks, there is a risk of disk group corruption for operations such as disk resizing. RESOLUTION: It is recommended that the FileStore cluster have more than one disk combined on all the storage pools. Etrack: 2729262 SYMPTOM: When a VIP fails over to other nodes, CIFS clients may experience connection issues to FileStore. DESCRIPTION: When a VIP fails over to other nodes, the Windows client still caches the old entry of the DFS referral, causing the Windows client to try and connect to the old VIP. The windows client may experience connection denied issues in this case. RESOLUTION: Reboot the windows client to clear the DFS cache. Etrack: 2735401 SYMPTOM: Adding mirror for nlm or DCO volumes can fail if there's only one pool with one disk on Filestore. DESCRIPTION: On Filestore, by default nlm and DCO volumes are mirrored for redundancy. When there's only one disk available, such operations can fail. Moreover, since the config copies of volume manager could not be distributed across multiple physical disks, there's a risk of disk group corruption for operations like disk resize. RESOLUTION: It is recommended that FileStore cluster should have more than 1 disk combined all the pools. Etrack: 2700193 Symptom: Missing information about support of Active Directory Japanese user names/groups from the Symantec FileStore documentation. Description: Symantec FileStore documentation should include information on support of Active Directory Japanese user names/groups. Resolution: Symantec FileStore supports Japanese, Korean, and Chinese local user/group names in areas supported by SAMBA. Etrack: 2753228 SYMPTOM: Supported version of Citrix XenDesktop missing from the FileStore 5.7 P1 README file. DESCRIPTION: With FileStore 5.7 P1, FileStore supports Citrix XenDesktop 5. The supported version of Citrix XenDesktop was not included in the FileStore 5.7 P1 README file. RESOLUTION: FileStore supports Citrix XenDesktop 5. Etrack: 2632963 SYMPTOM: tier relocate command will not relocate all files and tier remove command requires all its policy files to be removed. DESCRIPTION: Tier relocate will skip NDS files. NDS includes named data streams and extended attributes. Tier remove command will be successful only if all policies related to the tier are removed. man pages updated to reflect the same. RESOLUTION: There is no resolution for the NDS relocation issue. Run "tier policy remove " before running tier remove V. NEW FEATURES: ---------------- FileStore silent installation feature To use the FileStore silent installation feature, enter the following command: ./installer -f sfsinstall.conf An sfsinstall.conf.example file is provided in the FileStore tarball; it is located in the same directory as the installer. To use the sfsinstall.conf.example file, you need to do the following: 1. Rename the file to "sfsinstall.conf". 2. Modify the file by changing the cluster name and IP address ranges etc. Etrack incident: 2705826 A new tunable "fullspace" was added to the Storage> fs alert set/unset commands: Storage> fs alert set numinodes|numspace|fullspace|fullinodes value [fs_name,...] Storage> fs alert unset numinodes|numspace|fullspace|fullinodes [fs_name,...] fullspace is the tunable for setting an alert if the file system becomes full. For a file system to run efficiently, you should always reserve some space for the file system rather than using 100% of the space. When file system usage is above the limit set by the fullspace tunable, all the NFS/CIFS shares on the file system are automatically changed to read-only to prevent the file system from becoming full again. When you grow the file system or delete some files to free up space, the NFS/CIFS shares are automatically changed back to read-write (there might be a delay of up to five minutes) for the change to occur. By default, the fullspace tunable is set to 0, which means that the fullspace tunable is disabled. Etrack: 2712143 A new tunable "fullinodes" was added to the Storage> fs alert set/unset commands: Storage> fs alert set numinodes|numspace|fullspace|fullinodes value [fs_name,...] Storage> fs alert unset numinodes|numspace|fullspace|fullinodes [fs_name,...] fullinodes is a tunable for setting an alert if the file system becomes full. When inodes on a file system reach the limit of fullinodes, the NFS/CIFS shares on the file system are automatically changed to read-only. After the file system is changed to read-only, you need to delete some files from the file system, and a remount of the file system may be required for the NFS/CIFS shares to be changed to read-write. By default, the fullinodes tunable is set to 0, which indicates that the fullinodes tunable is disabled. Etrack: 2700164 CIFS segregated share feature In Symantec FileStore 5.7 P1, the CIFS export option "ip=virtualip" was removed and replaced with the sharename@virtual_ip full identifier. FileStore provides unified access to all shares through virtual IPs. With the FileStore CIFS segregated share feature, FileStore can provide isolation between file shares while the storage is hosted on the same physical cluster. The FileStore CIFS segregated share feature provides logical partitioning between the CIFS shares. The CIFS segregated share feature essentially binds the (virtual IP, share, file system) into one logical entity. If a CIFS share is added with the sharename@virtual_ip full identifier, the CIFS share is created by allowing only this virtual IP to access this CIFS share. CIFS> share show ShareName FileSystem ShareOptions share1@10.200.60.105 fs1 owner=root, group=root, fs_mode=1777, rw share1@10.200.60.106 fs2 owner=root, group=root, fs_mode=1777, rw If a CIFS share is added without the sharename@virtual_ip full identifier, the CIFS share is defined as a normal share, and all the virtual IPs can be used to access this CIFS share. FileStore CIFS segregated shares are hosted by the same SAMBA instance. Therefore, all clients should join the same domain or trusted domains. The clients cannot be in completely different domains. Within the same domain/trusted domains, different client pools can use the different and isolated shares. DEPRECATED FEATURE ------------------ Etrack incident: 2706509 Starting with the FileStore 5.7 P2 release, the following feature is no longer supported: * The CIFS> share modify sharename [cifsoptions] CLI command. DRIVER UPDATE ------------- Starting with FileStore Release 5.7, the FileStore software and the operating system are provided as two separate packages. FileStore no longer supports a driver update disk (DUD) for hardware. If the operating system cannot recognize your RAID controller for a local disk during installation, you need to visit your hardware vendor's Website and get the corresponding DUD. Symantec has updated the OS kernel for FileStore to SLES 2.6.32.43-0.4. If you encounter any driver issues, contact Symantec Technical Support.