IBM Storwize V7000 Unified ++++++++++++++++++++++++++ --------------------------------------------------------------------------------- CONTENTS 1. Introduction 2. Prerequisites 3. Code Levels 4. Problems Resolved and New Features 4.1 New features in IBM Storwize V7000 Unified File Module 1.4.3.5 4.2 APARs resolved in IBM Storwize V7000 Unified File Module 1.4.3.5 4.3 New features in Storwize V7000 7.2.0.9 4.4 APARs resolved in Storwize V7000 7.2.0.9 5. Further Documentation 6. Known Issues and Restrictions in this Level 7. Maximum Configurations 8. Supported Hardware 9. Pre-1.4.3.5 Release Information 9.1 Storwize V7000 Unified version 1.4.3.4 9.2 Storwize V7000 Unified version 1.4.3.3 9.3 Storwize V7000 Unified version 1.4.3.2 9.4 Storwize V7000 Unified version 1.4.3.1 9.5 Storwize V7000 Unified version 1.4.3.0 9.6 Storwize V7000 Unified version 1.4.2.1 9.7 Storwize V7000 Unified version 1.4.2.0 9.8 Storwize V7000 Unified version 1.4.1.1 9.9 Storwize V7000 Unified version 1.4.1.0 9.10 APARs resolved in previous Storwize V7000 7.2.x releases 9.11 APARs resolved in previous Storwize V7000 7.1.x releases ---------------------------------------------------------------------------------- 1. Introduction This document contains important information about version 1.4.3.5 of the IBM Storwize V7000 Unified software. ---------------------------------------------------------------------------------- 2. Prerequisites Software Upgrade Test Utility: The software upgrade test utility "upgradetest" must be run within two hours of starting the upgrade. The upgrade test utility can be run as many times as required, but there must be a successful run of the upgrade test utility within two hours of starting the upgrade. Refer to the following URL for further information: http://www.ibm.com/support/docview.wss?uid=ssg1S4000986 Concurrent Upgrades: The concurrent upgrade (upgrade with I/O running) restrictions are documented at the following URL: http://www.ibm.com/support/docview.wss?uid=ssg1S1004227#_Restrictions_Upgrade If you are performing a concurrent code upgrade, first ensure that all host paths are available and operating correctly. System Performance: Performance of the IBM Storwize V7000 Unified will be degraded during the software upgrade process. IBM recommends you perform the upgrade at a time of lower overall I/O activity and when significant spikes in the system workload are not expected. Event Checking: Check there are no unfixed events in the event log. Use normal service procedures to resolve any events before proceeding. Asynchronous Replication Systems: If the system is part of an asynchronous relationship (either source or target), stop the replication to avoid asynchronous replication failures during upgrade. Full upgrade compatibility matrix: Please see the following web page for the full upgrade compatibility matrix. http://www.ibm.com/support/docview.wss?uid=ssg1S1003918 ------------------------------------------------------------------------------------- 3. Code Levels V7000 Unified Software: V7000 Unified Product Version 1.4.3.5 (1.4.3.5-2) This includes Storwize V7000 (2076) Product Version: 7.2.0.9 (88.0.1410080000) ------------------------------------------------------------------------------------- 4. Problems Resolved and New Features 4.1 New features in IBM Storwize V7000 Unified File Module 1.4.3.5 None 4.2 APARs resolved in IBM Storwize V7000 Unified File Module 1.4.3.5 This Storwize V7000 Unified 1.4.3.5 release contains critical security fix for bash vulnerability known as "Bash Bug" or "Shellshock". It also contains other security fixes. For more information on the security fixes please refer to the flashes with "Security Bulletin:" in heading at below link (HIPER): http://www-947.ibm.com/support/entry/portal/alerts/hardware/system_storage/disk_systems/mid-range_disk_systems/ibm_storwize_v7000_unified 4.3 New features in Storwize V7000 7.2.0.9 None 4.4 APARs resolved in Storwize V7000 7.2.0.9 Bash vulnerabilities CVE-2014-6271, CVE-2014-7169, CVE-2014-7186, CVE-2014-7187, CVE-2014-6277, CVE-2014-6278 ----------------------------------------------------------------------------------- 5. Further Documentation The latest version of all IBM Storwize V7000 Unified documentation is available in the IBM Storwize V7000 Unified 1.4 Information Center: http://pic.dhe.ibm.com/infocenter/storwize/unified_ic/index.jsp Selected documents are available for download in PDF format from the support website: http://www.ibm.com/storage/support/storwize/V7000/unified Please refer to below link for a consolidated list of highly recommended reference material on IBM Storwize V7000 Unified http://www-01.ibm.com/support/docview.wss?uid=ssg1S1004460 ----------------------------------------------------------------------------------- 6. Known Issues and Restrictions in this Level * When creating a CIFS share the usage of %G or %g substitution variables inside the export path definition is not currently supported. This restriction will be lifted in a future IBM Storwize V7000 Unified release. * In IBM Storwize V7000 Unified, Windows 8 with offline folders is not currently supported. * Browsers with below listed levels are certified for using the IBM Storwize V7000 Unified Management GUI. Mozilla Firefox 23 Mozilla Firefox Extended Support Release (ESR) 17 Important: All Firefox users should upgrade to these updated packages, which contain Firefox version 17.0.7 ESR. After installing the update, Firefox must be restarted for the changes to take effect. For Red Hat Linux users, make sure all previously released errata relevant to your system have been applied before applying this update. This update is available at the Red Hat Network: https://rhn.redhat.com/errata/RHSA-2013-0981.html. Details on how to use the Red Hat Network to apply this update are available at https://access.redhat.com/knowledge/articles/11258 Microsoft Internet Explorer (IE) 8 and 9 Google Chrome 29 Later versions of the browsers are supported as long as the vendors do not remove or disable functionality that the product relies upon. For browser levels higher than the levels that have been certified with the product, customer support will accept usage-related and defect- related service requests. As with operating system and virtualization environments, if IBM support cannot recreate the issue in our lab, we may ask the client to recreate the problem on a certified browser version in order to determine if a product defect exists. Defects will not be accepted for cosmetic differences between browsers or browser versions that do not affect the functional behavior of the product. If a problem is identified in the product, defects will be accepted. If a problem is identified with the browser, IBM may investigate potential solutions or work-arounds that the client may implement until a permanent solution becomes available. * In IBM Storwize V7000 Unified, quarantined folders are only accessible with the root ID. This restriction will be lifted in a future IBM Storwize V7000 Unified release. Refer to the below technote for more information on this issue: http://www-01.ibm.com/support/docview.wss?uid=ssg1S1004078 For more details please refer to the V1.4 Configuration Limits and Restrictions for IBM Storwize V7000 Unified for product restrictions that currently apply: http://www.ibm.com/support/docview.wss?uid=ssg1S1004227 * Global Mirror Relationships Must Be Stopped When Upgrading to V7.2.0.x Refer to the following technote for more information on this restriction: http://www-01.ibm.com/support/docview.wss?uid=ssg1S1004472 * It is not supported to shrink the size of a compressed volume in Storwize V7000 V7.2.0.x * If using IP replication, please review the set of restrictions published in the V7.2 Configuration Limits and Restrictions document. * Moving Volumes to Another I/O Group Without Stopping Host Activity is Not Supported for All Operating Systems Refer to the following technote for more information on this restriction: http://www.ibm.com/support/docview.wss?uid=ssg1S1004125 * Intra-Cluster Global Mirror Not Supported Refer to the following flash for more information on this restriction: http://www.ibm.com/support/docview.wss?uid=ssg1S1003725 * Storwize Family Inter-cluster Metro Mirror and Global Mirror Compatibility Cross Reference Refer to the following flash for more information on this restriction: http://www.ibm.com/support/docview.wss?uid=ssg1S1003646 Once the software upgrade process has been started, the management GUI automatically refreshes to display the progress of the upgrade. During the upgrade each file module will be restarted, resulting in a fail-over and fail-back of the management services, including the GUI. After a fail-over has occurred it may be necessary to manually refresh/reload the page in the web browser in order to reconnect with the new active management node and login again. If an IBM Storwize V7000 Unified Code Upgrade Stalls or Fails then contact IBM Support for further assistance http://www.ibm.com/support/docview.wss?uid=ssg1S1002894 -------------------------------------------------------------------------------------- 7. Maximum Configurations The maximum configurations for IBM Storwize V7000 Unified 1.4.3.5 are documented in the V1.4 Configuration Limits and Restrictions for IBM Storwize V7000 Unified document which is available from: http://www.ibm.com/support/docview.wss?uid=ssg1S1004227#_MaximumConfigurations -------------------------------------------------------------------------------------- 8. Supported Hardware This release of IBM Storwize V7000 Unified software is supported on the following hardware types: File Modules: * 2073-700 * 2073-720 Storwize V7000 Control enclosures: * 2076-112 * 2076-124 * 2076-312 * 2076-324 Storwize V7000 Expansion enclosures: * 2076-212 * 2076-224 --------------------------------------------------------------------------------------- 9. Pre-1.4.3.5 Release Information 9.1 Storwize V7000 Unified version 1.4.3.4 Code Levels: Storwize V7000 Unified Product Version 1.4.3.4 (1.4.3.4-6) Storwize V7000 (2076) Product Version 7.2.0.8(88.0.1408150000) New features in IBM Storwize V7000 Unified File Module 1.4.3.4 None APARs resolved in IBM Storwize V7000 Unified File Module 1.4.3.4 Critical Fixes HU00479 Upgrade from Storwize V7000 Unified V1.4.3.2 fails with error code 0520:Error on restore of database. This can be seen when an upgrade is restarted after an earlier upgrade failure. This Storwize V7000 Unified 1.4.3.4 release also contains some critical security fixes For more information on the security fixes please refer to the flashes with "Security Bulletin:" in heading at below link: http://www-947.ibm.com/support/entry/portal/alerts/hardware/system_storage/disk_systems/mid-range_disk_systems/ibm_storwize_v7000_unified 9.2 Storwize V7000 Unified version 1.4.3.3 Code Levels: Storwize V7000 Unified Product Version 1.4.3.3 (1.4.3.3-11) Storwize V7000 (2076) Product Version 7.2.0.7 (87.9.1406171000) In IBM Storwize V7000 Unified V1.4.2.x, V1.4.3.0, V1.4.3.1, V1.4.3.2 releases there is a known GPFS issue which can cause undetected or silent data corruption when the file modules are updating a shared directory concurrently. This issue has been fixed in this release. Refer to the below flash for more information on this issue: http://www-01.ibm.com/support/docview.wss?uid=ssg1S1004806 New features in IBM Storwize V7000 Unified File Module 1.4.3.3 None APARs resolved in IBM Storwize V7000 Unified File Module 1.4.3.3 Critical Fixes HU00433 V7000 Unified CLI command 'chuser' and 'mkuser' allow users to enter plain-text password as part of the command argument and this leaves the password visible in bash_history. And as bash_history is collected as part of the support package this causes a security violation HU00441 Due to incorrect NTP configuration settings the V7000 Unified system becomes exposed to security vulnerability CVE-2013-5211 High Importance Fixes HU00435 On removing all the ACL entries from a share the Export Owner ends up loosing access to the share HU00458 CTDB on a file module becomes unhealthy and the var/log/messages on the file module continuously report nfs_knfsd errors and requires reboot to recover. This happens due to a Linux kernel bug causing the NFS threads to deadlock HU00437 Network failure detected after removing a network interface bond and then creating it with a different MTU (maximum transmission unit) than the default This issue is mostly seen when changing MTU on non-management network interface bonds HU00439 File module crashes and reboots when there are simultaneous reads on a single large file from multiple NFS clients This Storwize V7000 Unified 1.4.3.3 release also contains some critical security fixes For more information on the security fixes please refer to the flashes with "Security Bulletin:" in heading at below link: http://www-947.ibm.com/support/entry/portal/alerts/hardware/system_storage/disk_systems/mid-range_disk_systems/ibm_storwize_v7000_unified 9.3 Storwize V7000 Unified version 1.4.3.2 Code Levels: Storwize V7000 Unified Product Version 1.4.3.2 (1.4.3.2-4) Storwize V7000 (2076) Product Version 7.2.0.31 (87.7.1404220000) Note: Storwize V7000 V7.2.0.31 is only available as part of Storwize V7000 Unified V1.4.3.2. It is not intended for use on a Storwize V7000 block system when not part of a Storwize V7000 Unified. Storwize V7000 V7.2.0.31 included in this release is equivalent to Storwize V7000 V7.2.0.3 with additional fixes listed in the section 9.8 Issues resolved in IBM Storwize V7000 Unified File Module 1.4.3.2: OpenSSL Heartbleed vulnerability (CVE-2014-0160) After upgrading to V1.4.3.2 there are a few more steps that needs to be run to remediate the OpenSSL Heartbleed vulnerability. Please refer to the below link for the additional steps and information: http://www-01.ibm.com/support/docview.wss?rs=591&uid=ssg1S1004616 9.4 Storwize V7000 Unified version 1.4.3.1 Code Levels: Storwize V7000 Unified Product Version 1.4.3.1 (1.4.3.1-4) Storwize V7000 (2076) Product Version 7.2.0.3 (87.4.1402180000) New features in IBM Storwize V7000 Unified File Module 1.4.3.1: None APARs resolved in IBM Storwize V7000 Unified File Module 1.4.3.1: Critical Fixes HU00221 GUI fails to load in browser for non-english locale as preferred language 9.5 Storwize V7000 Unified version 1.4.3.0 Code Levels: Storwize V7000 Unified Product Version 1.4.3.0 (1.4.3.0-17) Storwize V7000 (2076) Product Version 7.2.0.3 (87.4.1402180000) When upgrading from IBM Storwize V7000 Unified V1.4.1.x or later versions the software upgrade test utility might give the below error: ============================= ERROR:NSD volumes are not mapped correctly. Refer to http://www.ibm.com/support/docview.wss?uid=ssg1S1004483 for more info. ============================= If you get this error your system is exposed to a potential data loss scenario. To correct this mismapping and avoid the potential data loss, please implement the action plan given in the above link. New features in IBM Storwize V7000 Unified File Module 1.4.3.0: * Increased the default number of NSDs (Network Shared Disks) created for a file system from GUI * Added functionality in 'chdisk' command to change the designation of a disk which is assigned to a file system * Added feature to run the Storwize V7000 upgradetest utility when doing a first time install of Storwize V7000 Unified * Disabled the Linux THP (Transparent Huge Pages) memory management function APARs resolved in IBM Storwize V7000 Unified File Module 1.4.3.0: Critical Fixes HU00194 Upgrade from V1.4.1.x to V1.4.2.x fails when there are stale NSDs on the system High Importance Fixes HU00171 CTDB unnecessarily restarts NFS service after a node failover. And this unnecessary restart of NFS service when other background problem exist results in NFS failing to start properly HU00199 On creation of a file system the paths to the newly added storage devices are not managed properly, leading to I/O getting blocked on these devices HU00186 While updating user quota setting from management GUI, the GUI browser window hangs HU00190 Strong warning to mention the impact on performance is not given when refreshing the usage of snapshots HU00202 NDMP failure with many filesystems registered HU00163 When the number of files on a share goes beyond the 32-bit inode range, into the 64-bit range, the 32-bit clients attached to the share lose access HU00208 Setting of new quotas on old or new filesets fails after lots of quota related activities and fails to recover on its own. Once this error occurs all quota related activities fail for that fileset HU00206 Node hangs during failover causing NFS clients to hang until node is rebooted HU00198 Severe performance degradation happens when doing asynchronous replication And this is due to sparse file optimization introduced in V1.4.1.0 HU00218 when upgrading the system from V1.3.x to V1.4.2.0 the 10Gb Emulex CNA adapter fails to upgrade to the latest firmware level HU00216 TSM restore becomes very slow for a HSM managed file system having RESTOREMIGSTATE set to YES (default setting of TSM client) Suggested Fixes HU00200 Event log shows a false warning message of failed chkpolicy even when chkpolicy completes successfully. This happens if the file system name contains an underscore character HU00162 TSM restore with multi-threading capacity is not allowed HU00211 ACLs cannot be set on file using the EMC documentum application HU00212 Need update in V7000 Unified information center on the page NFS considerations and limitations for the "The AIX NFS client has a know issue in retransmission behavior" line. 9.6 Storwize V7000 Unified version 1.4.2.1 Code Levels: Storwize V7000 Unified Product Version 1.4.2.1 (1.4.2.1-8) Storwize V7000 (2076) Product Version 7.1.0.6 (80.4.1310310000) When upgrading from IBM Storwize V7000 Unified V1.4.1.x or later versions the software upgrade test utility might give the below warning: ============================= STDOUT WARNING: One or more file system disks (a.k.a Network Shared Disks or NSDs) are not properly configured. It is strongly recommended you schedule a non-concurrent maintenance window to have the NSDs reconfigured as soon as feasible to avoid any access impacts that may be triggered due to this exposure, especially under certain file module failover conditions. Refer to http://www.ibm.com/support/docview.wss?uid=ssg1S1004483 for more details. ============================= If you get this warning then IBM support strongly recommends to implement the action plan given in the above link to correct this misconfiguration. Refer to http://www.ibm.com/support/docview.wss?uid=ssg1S1004483 for more details. New features in IBM Storwize V7000 Unified File Module 1.4.2.1: None APARs resolved in IBM Storwize V7000 Unified File Module 1.4.2.1: Critical Fixes HU00158 cnscm stops the affected node when GPFS long waiters are found High Importance Fixes HU00150 When the backup job is big TSM backup creates temporary files that fill up the root (/) partition making the node unhealthy HU00152 NFS client gets ESTALE error or hangs when netgroups are used and the cumulative name length of the exports is higher than 512 bytes HU00115 The upgrade test utility does not recognize the new 400GB SSD drives HU00116 V7000 Unified GUI does not report the used capacity of snapshots in the overall used capacity HU00124 GPFS policy errors do not propagate to the NDMP application HU00146 The core dumps generated by java crashes fill up the root (/) filesystem HU00120 'mmdefedquota' when run against a large number of users causes GPFS daemon to crash HU00135 The M4 servers in V7000 Unified have misconfiguration for PXE boot. Due to this in a PXE boot environment, the file module tries boot from network after reboot, and fails to boot up HU00143 Due to a big job list the 'lsjobstatus' runs out of memory and crashes HU00154 Callhome process reboots the file module when transfer of data hangs Suggested Fixes HU00139 'startrestore' man page has the wrong description for '-t' option HU00144 Can see the error "tail: write error: Broken pipe" in the TSM backup log This Storwize V7000 Unified 1.4.2.1 release also contains some critical security fixes For more information on the security fixes please refer to the below link: http://www-947.ibm.com/support/entry/portal/alerts/hardware/system_storage/disk_systems/mid-range_disk_systems/ibm_storwize_v7000_unified 9.7 Storwize V7000 Unified version 1.4.2.0 Code Levels: Storwize V7000 Unified Product Version 1.4.2.0 (1.4.2.0-27) Storwize V7000 (2076) Product Version 7.1.0.5 (80.4.1309270000) New features in IBM Storwize V7000 Unified File Module 1.4.2.0: * Automatically exclude migrated files with asynchronous replication * Increased the number of supported CIFS exports to 10000 * Added feature to enforce mount state with active exports * Management Access recoverable after a network reconfiguration * Added feature to disable CIFS IO on share deactivation or removal * Added support for 'AD + LDAP' authentication and Enhanced Firewall rules * Call home installed by default This Storwize V7000 Unified 1.4.2.0 release also contains some critical security fixes For more information on the security fixes please refer to the below link: http://www-947.ibm.com/support/entry/portal/alerts/hardware/system_storage/disk_systems/mid-range_disk_systems/ibm_storwize_v7000_unified APARs resolved in IBM Storwize V7000 Unified File Module 1.4.2.0: Critical Fixes HU00033 GPFS assert causes node to crash HU00066 System crashes when collecting network trace due to CTDB on all nodes getting Banned at the same time HU00106 CTDB goes into Banned state on many nodes which leads to an outage HU00073 GPFS DirectoryBlock split waiters lead to an outage HU00010 Access loss seen after deleting a large fileset HU00067 Fileset operations such as snapshot deletion or inode space expansion on a heavily loaded systems lead to GPFS deadlock causing the system to hang High Importance Fixes HU00006 Asynchronous replication settings get removed from the data base causing the replication to fail HU00053 CTDB is banned on one of the node and this consumes a lot of CPU HU00043 Multiple GPFS waiters hang the node HU00023 Callhome entries longer than 250 characters are not sent to IBM HU00037 CTDB is not marked Unhealthy when the nfs service is unresponsive on the node HU00039 High I/O load causes memory leak in the QLogic driver HU00064 Performance issues seen with the Emulex firmware/driver HU00065 Error recovery in asynchronous replication will recover only failed files instead of all the failed split files HU00074 CNSCM makes the node healthy even when the GPFS filesytems are not mounted leading to stale NFS handles HU00095 'startbackup' fails and returns wrong error code HU00088 snmptrapd process consumes 100% CPU leading to performance issues HU00093 After NFS is restarted the NFS thread count becomes less than the configured value HU00032 Getting the error "Could not get info for group" when resolving CIFS groups HU00079 Asynchronous replication causes performance degradation after upgrading to 1.4.1.0 HU00083 Few NFS options are reset to default when editing a NFS share from the GUI HU00025 Customer getting the error "EFSSA0156W - A problem occurred in VPD task on host mgmt001st001" after upgrading to 1.4.1.0 HU00040 CTDB stays Unhealthy on a node as it is unable to access the CNSCM port HU00044 1.4.0.x to 1.4.1.x upgrade test utility failed with version mismatch error HU00059 Asynchronous replication fails as unable to create rsync temporary files HU00103 Lost access to NFS shares on rebooting the V7000 canisters HU00058 Node crashes due to load and insufficient free memory HU00098 Changing the windows read_only attribute also changes the file permissions to 744 from 444 HU00084 Maximum allowed concurrent backup per cluster is 1 HU00086 Unable to mount GPFS filesystem on mgmt001st001 due to changes in /etc/fstab HU00076 On running the software upgrade test utility the management service is always restarted HU00100 'ctdb rebalancenode' command causes IPs to be assigned to Unhealthy nodes Suggested Fixes HU00012 Deleting a running asynchronous replication session leads to false entries in GUI HU00024 CIFS locking does not work using Fedora Linux and KDE desktop and CIFS client login does not work with SMB1 and KDE desktop IC96118 After adding an email recipient and enabling email notification the GUI throws an HTTP error when accessed HU00063 Single "dd" command from client with options bs=1024k and count=5120 gives slow performance HU00075 SONAS software upgrade test utility does not check IMM/ASU connectivity on all the nodes HU00091 Need first time data capture improvement for adding /proc/lock output from all the nodes into cndump HU00035 Customer cannot save files with extension .xlsx on SONAS CIFS shares HU00081 Few critical error messages in the pre-upgrade checks have confusing message HU00005 InfoCenter does not reflect the hardware design for 2073-720 HU00048 First time installation error message "0BA4 Unable to attach to Storwize Controller" is misleading and incorrect HU00096 In some cases 'cnlsregistry' creates unnecessary files in /tmp files which leads to filling up of the /root filesystem HU00097 The tooltip on a mapped share always shows IBM Sonas even if the volume is from Storwize V7000 Unified HU00046 HP Webinspect flags security vulnerability with V7000 Unified HU00061 When a user on a Windows client lists the available snapshots on a V7000 Unified share through the "Previous Versions" dialog, the displayed data is not as expected. The dialog shows entries starting with @GMT instead of the subdirectories in the snapshot 9.8 Storwize V7000 Unified version 1.4.1.1 Code Levels: Storwize V7000 Unified Product Version 1.4.1.1 (1.4.1.1-4) Storwize V7000 (2076) Product Version 7.1.0.2 (79.8.1307111000) New features in IBM Storwize V7000 Unified File Module 1.4.1.1: None APARs resolved in IBM Storwize V7000 Unified File Module 1.4.1.1: Critical Fixes IC94791 CTDB getting banned on all nodes at the same time causes an outage High Importance Fixes IC94750 A typo in mmbackup causes the policy command invoked by the backup scripts to use /tmp for its temporary storage rather than the directory provided to the backup scripts on the command line IC94876 CTDB does not monitor for missing public network interface and remains healthy on the node where a network interface with public IPs configured is missing IC94798 NDMP incremental backup fails if the directory path name contains space IC94792 V7000 Unified does not correctly respond to CIFS client GetInfo request when output buffer length is too small Suggested Fixes IC94800 Callhome performs many FTP connections to ecurep.ibm.com if a file to be uploaded does not exist on the V7000 Unified machine 9.9 Storwize V7000 Unified version 1.4.1.0 Code Levels: Storwize V7000 Unified Product Version 1.4.1.0 (1.4.1.0-40) Storwize V7000 (2076) Product Version 7.1.0.1 (79.6.1306101000) IBM Storwize V7000 Unified systems that are utilizing compressed volumes will be prevented from upgrading to this Storwize V7000 Unified V1.4.1.0 release. This restriction will be lifted with a future release of Storwize V7000 Unified. Refer to the following flash for more information on this restriction: http://www.ibm.com/support/docview.wss?uid=ssg1S1004393 New features in IBM Storwize V7000 Unified File Module 1.4.1.0: * Added support for Active Cloud Engine global file sharing between IBM SONAS and IBM Storwize V7000 Unified APARs resolved in IBM Storwize V7000 Unified File Module 1.4.1.0: Critical Fixes IC93380 ctdb.event files generated repeatedly causing /dev/sda0 to reach 100% IC93386 Upgrade to V1.4.0.1 failed on node mgmt001st001 with error code 0517 IC93388 Service IPs overwritten by using an USB drive with changed configuration files IC93392 Incorrect check for empty event list from V7000 lseventlog output causes upgrade to fail IC93395 Upgrade process stopped with error code 0521 while upgrading to V1.4.0.1 due to a race condition when switching from active to passive management node IC93451 Upgrade to V1.4.0.1 failed with error code 01A8 IC93462 mkfs fails creating filesystem with error code EFSSG0476C This will mostly occur when there are orphan sd devices (sd devices not belonging to a mpath or vdisk device) on the system IC93463 V7000 Unified upgrade issues IC93485 During high load conditions upgrade fails with error "0521,U145node_sync_database,096" IC93486 Overlapping network configurations led to access loss IC93517 Message EFSSF1217C was detected during Upgrade to V1.4.0.1 IC93587 CTDB not working properly results in access loss IC93685 Upgrade to V1.3.2.3-8A failed with 01BA during HSM upgrade IC93687 Upgrade checks do not reveal non-redundant network configuration IC93693 After upgrade system reported exception in the gpfssnapshot gatherer task IC93439 GPFS asserts on removing suspended NSD using mmdeldisk command High Importance Fixes IC93371 chnwmgt does not update /etc/ssh/sshd_mgmt_conf file on passive management node IC93383 After upgrade to V1.4.0.1 the V7000 Unified GUI disallows logins IC93389 Performance graphs did not populate with any data as expected IC93391 /ftdc folder becomes full IC93393 Write access to a file in a snapshot via CIFS share possible IC93397 Removing a file system from the cluster does not remove the mount point on the passive management node IC93398 LDAP authentication issues after upgrading to V1.4.0.1 IC93400 CPU threshold exceeds time to time and CIM reports mgmt_switchover and python processes as cause IC93402 CIM agent is crashing and deleting core dumps IC93403 Async replication after a management switchover was full sync while a replication was not running during management switchover IC93404 Long load times when browsing for 'path' on the create export GUI panel IC93417 ILM migration severely impacted performance IC93423 CTDB IPs conflict with file module Service IPs IC93434 Files from /root of the passive management node disappear after management node replication IC93465 Async replication was stopped due to exceeding the threshold for files to recover IC93471 Error message in the upgrade test utility 1.4.0.x-3 IC93479 Long delays on node reboot during upgrade IC93481 Due to stale information related to V7000 HDD firmware levels software upgrade test utility 4.4 throws warnings IC93506 Modification to the GPFS CIFS oplock for file stamp modification to prevent lease break and performance issues IC93507 Some applications can fail because /var directory filling up IC93538 Performance degradation caused by ctdb record contention IC93562 Filesystems with underscore in name not handled properly during backup IC93567 SMB/SMB2 client connections are unstable when SMB signing is enforced via Windows security policies IC93588 cndump process runs in the background even after finishing data collection IC93590 Issues with V7000 Unified GUI IC93592 Wrong IPs displayed on V7000 Unified GUI IC93660 After upgrading from V1.3.x to V1.4.0.x performance Center does not start IC93696 Replication of filesystem fails due to memory corruption IC93699 Callhome unable to communicate using proxy servers after upgrade IC93515 /var file system has been filled up with winbindd logs and messages IC93566 SI0100W : GPFS Timeout - mmgetstate did not respond within 20s IC93568 Internal processes can increase over a period of time and affect performance IC93701 chnode command results in GPFS going down and nodes rebooting IC93713 Async replication log gathering takes long time IC93714 Async replication delete phase takes long time with large number of files IC93753 An unexpected exception occured while running a gatherer task IC93716 Call home not working after upgrade IC93752 Login to GUI is not possible after management switchover IC93717 ntpd service stopped is not reported in syslog IC93742 Cache pre-population fails IC93757 GUI reports SI0126W event for passive management node IC93719 Misleading events given during management node failover or restart IC93412 PMR not generated for CallHomes IC93721 Changes to filesystem cause events with NSD thresholds IC93722 The used capacity for filesets and filesystems is not the aggregated capacity of the snapshots IC93422 CTDB wrongly showing as unhealthy on all nodes IC93427 No indication in GUI when no free space left for metadata IC93438 NFS configuration via GUI does not follow the --nfsPosition option IC93579 Backup error logs have incorrect timestamps IC93443 Due to async incorrectly updating its internal status file the retry after the error does not run rsync phase IC93581 Warning event "The specified command append2 could not be composed for OS RHEL 6.1 x86_64" gets logged periodically IC93453 Independent filesets cannot be created if the filesystem has already reached the maximum number of inodes allowed IC93455 Switching over management node leads to postgres database corruption IC93458 CTDB IPs are not rebalanced ater a quick healthy-unhealthy-healthy cycle on a node IC93470 Rsync fails to sync a specific file and exit with RC=12 IC93488 The system clock on individual nodes is drifting in different directions IC93490 lsjobstatus for a backup job shows failure due to skipped socket files IC93496 HSM does not shutdown on suspended node IC93500 Unclear status reported by backup operation IC93723 System is logging the event id SL0104W indicating HSM is not configured IC93724 Message indicating CTDB is stopped is displayed on lslog even if CTDB is running IC93712 Message "The passive management node replication status changed to DATABASE REPLICATION FAILED." found in event log IC93711 Auto-migration task running on filesystems even with free capacity IC93710 On systems with a lot of snapshots, lssnapshot -u (usage) takes a long time to complete IC93709 runpolicy command can not be run while backup is running IC93708 Erroneous full name returned by winbind IC93603 Customer getting error code "sg0002c" IC93622 Need clarification on perserving case with file systems and share names IC93607 CIFS users not able to list files on CIFS exports IC93611 GPFS ACL garbage collector has problem with ACLs and independent filesets IC93615 System unable to determine the current CIM server status IC93621 GPFS service restarts on management node IC93624 As seen by CTDB a node is healthy but does not serve SMB anymore IC93628 CTDB logging the following message repeatedly: 'Vacuuming child process timed out' IC93630 CTDB IP takeover on management node did not complete successfully IC93634 FTP does not work correctly with Windows explorer or MAC finder IC93649 10G interface (ethX1) does not come up after nodes recover from crash IC93652 Users see false positive MMFS_FSSTRUCT error appear in the system log with error code of 1124 IC93656 %U option listed in the Information Center is not accepted in GUI when creating shares Suggested Fixes IC93381 Incorrect description for replacement of a hot-swap drive in a file node in the V7000 Unified Information Center IC93382 V7000 Unified Information Center documentation issue for hot swap drive IC93430 Temporary loss in connection with Windows Domain Controller gives errors IC93447 Commands which ask the user "Do you really want?" (mostly rm commands) return with a exit code of 0 if the user answers "no". This makes it impossible to distinguish between run and aborted commands in the audit log. IC93512 The GUI and lsnwntp show 127.127.1.1 in the ntp server list Which is not necessary to be shown IC93513 Uncompleted documentation in the information center for "chservice" command IC93532 Unwanted error message "TAIL: WRITE ERROR: BROKEN PIPE" shown during backup IC93533 lshsm prints a blank line in the output IC93557 Truncated trace data collected by gpfs.snap --deadlock IC93564 Debug data not captured when CTDB fails to freeze databases as part of recovery IC93706 Undeclared message ID from subsensor NSLCD in lshealth reported IC93705 System cnrssccheck timeouts IC93703 mkclone command not able to create multiple clones of same parent IC93704 ILM policies do not use fileset name to create policy in GUI IC93700 To help development in analysing TSM/HSM related issues GPFS related information (mmdrfs) is needed in cndump IC93702 Time stamp data when refreshed is not changing for list of filesets in GUI IC93765 Call home record contains only serial numbers of active management node IC93733 Confusing mail notification regarding backup jobs SA0187W IC93763 srvdump does not allow operating on dump IDs having commas IC93762 When using RAS menus to offload the data to IBM via call home no proper information given on how to specify a PMR IC93759 No confirmation messages given when using RAS menu to shutdown entire V7000 Unified IC93758 Add stack tracing for certain Java errors IC93449 Maximum email sent per hour setting does not work IC93736 Multiple test call homes generated IC93755 The message presented using "srvdump -d" command is not clear to indicate that the maximum number of automatic and system generated dumps are reached IC93754 When displaying policies using the GUI customer found the following issue in the GUI: Carriage return are lost IC93695 Callhomes not allowed to use hostnames as proxy if they are not able to ping during cnrscallhomeconfig IC93619 Callhome message is not clear 9.10 APARs resolved in previous Storwize V7000 7.2.x releases: 7.2.0.8 Critical Fixes HU00280 Multiple node asserts triggered by a Global Mirror disconnection. HU00283 Multiple node asserts caused by invalid compressed volume metadata. HU00450 Manual upgrade with stopped Global Mirror relationships can not complete non-disruptively due to dependent volumes. High Importance Fixes HU00176 Node assert dies to an IO deadlock when using the remote copy functions. HU00253 Global Mirror with Change Volumes does not resume copying after an IO group goes offline at secondary cluster. HU00264 Node assert caused by an AIX host logout colliding with a new SCSI check condition being generated. HU00274 Quiesce and resume of host IO when Global Mirror consistency group reaches consistent_synchronized. HU00286 Node assert due to compression running out of resources. HU00293 Node canister fails to boot after a hard shutdown. IC90374 Node assert due to a IO deadlock when using the FlashCopy functions. Suggested Fixes HU00109 Fix procedure for 1920 Global Mirror event fails to resolve the issue. HU00134 Stopped Global Mirror Change Volumes relationship cannot be added to a stopped consistency group. HU00223 Fix procedure for 1685 failed to leave drives in correct state. HU00241 GUI runs out of resources to issue commands to the system. HU00252 Service Assistant GUI fails to run commands with a message that the user is not logged in as superuser. HU00272 RAID resync progress incorrectly reported as 0% after upgrade. 7.2.0.7 New features in Storwize V7000 7.2.0.7 * Increased write throughput due to full stride write enhancements * SCSI WriteSame to drives has been disabled to improve performance of writing zeros to drives. Note this may slow down initialisation of arrays under some scenarios. This does not affect SCSI WriteSame from hosts to the system. OpenSSL vulnerability CVE-2014-0224 Critical Fixes HU00126 Multiple node asserts after node failures if using thin provisioning HU00191 Multiple node asserts if iSCSI alias contains one or more spaces HU00245 Multiple node asserts when secondary system running 7.2.0 and systems become disconnected HU00275 Multiple node asserts when secondary system running 7.2.0 and systems become disconnected High Importance Fixes HU00348 Offline MDiskgroup when using Thin Provisioning HU00226 Offline compressed volume caused by fault during automatic space reclamation HU00254 Single node assert when two systems disconnect from each other HU00270 Unnecessary delays during node restarts Suggested Fixes HU00027 iSCSI performance degradation with large application I/O operations HU00175 Performance degradation when cache gets stuck at 80% full HU00242 Host workloads can generate a large amount of compression metadata on disk, reducing compression savings HU00246 Global Mirror Secondary Write Lag performance metric reported in microseconds instead of milliseconds on 7.2.0 secondary clusters HU00258 Global Mirror Change Volumes on secondary system do not automatically shrink when running V7.2.0 HU00260 Global Mirror internal tracing insufficient for problem determination in certain scenarios HU00256 Invalid output on STDERR causing TPC probes to fail on V7.2.0.0 or later 7.2.0.31 Issues resolved in Storwize V7000 7.2.0.31: OpenSSL vulnerabilities (CVE-2014-0160 and CVE-2014-0076) 7.2.0.3 New features in Storwize V7000 7.2.0.3: * Multi-drive firmware download * Configurable port masking * Global Mirror performance improvements * Support for remote copy replication using native IP links * Multi-language support Critical Fixes HU00178 Loss of access to data when expanding volume size of an offline compressed volume HU00230 Loss of access to data when stopping Metro/Global Mirror relationships on 7.2.0.1 HU00014 Loss of access to data in systems with a large number of volumes mapped to hosts HU00019 Multiple node asserts when inter-canister link can not be initialised HU00126 Multiple node asserts when all nodes containing FC metadata are offline, but volumes are online IC88865 Node asserts caused by handling excessive number of faults from backend storage IC90476 Node asserts if the ldap server with id 0 is not configured IC91332 No configuration node after upgrading with GMCV active IC92665 Node assert when an iSCSI initiator has the same IQN as the iSCSI target IC92867 Node asserts caused by moving the only compressed volume in an IO group to a different IO group High Importance Fixes HU00220 Security issue CVE-2014-0880 HU00161 Daily configuration backup not being distributed to non configuration nodes HU00167 Node assert caused by premature end of decompression activity HU00229 Node assert caused by lock contention HU00027 iSCSI performance instability for large transfers HU00132 Node assert when repairing killsectors on RAID 10 arrays HU00167 Node assert due to decompression issue IC87264 Node assert caused by timing window shortly after starting an incremental FlashCopy IC89608 Node assert caused by XCOPY command to a recently offline volume IC91336 Unable to detect Storwize LUNs when connecting backend storage using FCoE IC94392 System was returning different messages for invalid username and invalid password IC95107 Node assert when large numbers of space efficient volumes come online at the same time on a single node Suggested Fixes HU00069 IERR log file incorrectly generated on power loss HU00108 GUI incorrectly warns that certain ldap username strings are invalid HU00110 GUI issues command without details in progress dialog HU00112 GUI fails to complete operations, with no way to close the task dialog box HU00127 Incorrect progress information for Global Mirror with Change Volumes HU00169 Too many error log dumps stored on a node IC86836 Event ID 45040 if battery maintenance is ocurring IC92993 Drive replacement DMP failed to fix drive that was dead on arrival IC93080 Unexpected enclosure in system after Cluster Recovery (T3) procedure IC94781 Health status pod not auto-updating when issues are resolved 9.11 APARs resolved in previous Storwize V7000 7.1.x releases: 7.1.0.6 Critical Fixes IC92390 Node asserts when a node asserts whilst processing multiple SCSI CAW commands HU00016 Node error 578 caused by a large number of failed SSH connections on 7.1.0 High Importance Fixes HU00121 Node assert due to memory issues when using compressed volumes HU00123 Compressed Volume offline after running repairsevdiskcopy Suggested Fixes IC89214 Fibre channel logins fail to establish between clusters due to dropped FC frames 7.1.0.5 Critical Fixes HU00002 Multiple node asserts during upgrades when using VMware hosts with VAAI HU00013 Loss of access to data during routine 90-day battery maintenance cycles on batteries shipped after March 2013 HU00014 Multiple repeated node asserts when large numbers of volume-to-host mappings are created HU00018 Node asserts when FlashCopy bitmaps reside in different IO group than FlashCopy source volumes HU00021 IO group-wide performance impacts when receiving large numbers of aborts from hosts 7.1.0.3 Critical Fixes IC93153 Loss of host access caused by mishandling of SCSI Persistent reserves during upgrades IC94405 Global Mirror Change Volume target not correctly updated with source data when relationship is stopped if the source volume is in any FlashCopy maps other than those managed by Global Mirror High Importance Fixes IC92386 Node assert when software upgrade completes on system with mirrored thin provisioned volumes IC94566 Multiple node asserts when RAID member exchange encounters a medium error which is greater than 2 TB from the start of the drive IC94935 Node assert when an I/O is sent beyond the end of a compressed volume IC94937 Node assert due to timing window when using compression 7.1.0.2 Critical Fixes IC91058 Multiple node asserts when fixing an event log entry if more than 10 event log entries have the root sequence number of the event being fixed IC91425 Multiple node asserts when moving a volume that is the target of a stopping FlashCopy map to another IO group Suggested Fixes IC90824 Configuration backup could only be run by superuser IC94125 Drive with product id MK3001GRRB failing unnecessarily 7.1.0.1 New Feature * Ability to configure FC ports for specific traffic types Critical Fixes IC93087 Loss of access to data due to internal bus error IC83684 Hardware fault in one node canister causes reboot of both node canisters IC86864 Node canister shutdown during battery firmware upgrade IC88865 Loss of access to data when system is logging a very large number of errors about backend storage IC90295 Node warmstarts during migrations caused by persistent reservations on image mode managed disks IC90476 Loss of access to data when an LDAP server 0 is not configured IC90637 Node warmstarts caused by RNID frames sent to FCoE ports from SAN management tool IC90872 Multiple node warmstarts caused by slow I/O from backend storage IC91255 Using two or more FlashCopy maps with the same target volume may result in inconsistent data on the target volume High Importance Fixes IC83241 Single node warmstart due to delays processing Global Mirror IOs IC86399 Single node warmstart when significantly increasing the size of a volume IC86591 Slow response time when host fails to send data after system sends XFER_RDY IC86971 Global Mirror slowdown when host fails to send data after system sends XFER_RDY IC87048 Single node warmstart due to inter-cluster communication issues IC88869 iSCSI Performance issues IC90060 Single node warmstart caused by timing window in EasyTier IC90442 Single node warmstart caused by unexpected settings in FCP frame IC90821 Single node warmstart caused by frequent GUI logins IC92028 Cache not sending enough IOs to compression IC92138 Single node assert when force deleting a volume involved in FlashCopy maps IC92386 Single node warmstart during software upgarde commit due to slow IO response times for Thin Provisioned volumes IC92803 Single node assert caused by SCSI Compare and Write command Suggested Fixes IC93080 Additional enclosure is seen after cluster recovery procedure IC93081 Enclosure becomes temporarily degraded IC93083 Flash Copy exceeds configured background copy rate IC93088 1625 errors when removing unmanaged mdisk IC84694 Incorrect Fix Procedure for 1840 event IC79537 Invalid 1048 event logged when canister is offline IC82729 1692 Fix Procedure not marking event as fixed IC86279 Incorrect title for Edit Port panel in iSCSI GUI IC86505 Fibre Channel ports will not function if connected to a switch port with a Domain ID ending in 'FF' IC87229 GUI shows empty tables when CPU utilisation is high IC87339 GUI dialog hangs at Synchronizing memory cache after changing enclosure ID IC88220 Slow boot time and possible node warmstart during the detection of ethernet ports IC89258 Array members constantly exchanging after changing enclosure ID IC89378 DS3000, DS4000 and DS5000 managed disks not using preferred controller IC89661 Unable to create snap package if too many files exist IC90771 Drive replacement fix procedure does not remove old drive IC91142 Temporarily unable to create FlashCopy of Global Mirror Change Volume Secondary volume after stopping the Global Mirror IC91336 Intel FCoE software initiator fails to detect LUNS IC91993 GUI shows wrong Metro/Global Mirror direction when stopped IC92075 Single node assert due to resource leak in FCoE driver ----------------------------------------------------------------------------------------------------