Patch-ID# 121069-01 NOTE: *********************************************************************** READ THE TERMS OF THE AGREEMENT ("AGREEMENT") IN THE LEGAL_LICENSE.TXT FILE CAREFULLY BEFORE USING THIS SOFTWARE. BY USING THE SOFTWARE, YOU AGREE TO THE TERMS OF THIS AGREEMENT. IF YOU DO NOT AGREE TO ALL OF THE TERMS, PROMPTLY DESTROY THE UNUSED SOFTWARE. *********************************************************************** Keywords: esm enterprise storage manager advanced applications Synopsis: Sun StorEdge ESM 4.0 Advanced Applications Patch Date: Oct/21/2005 Install Requirements: None Solaris Release: 8 9 SunOS Release: 5.8 5.9 Unbundled Product: StorEdge ESM Advanced Applications Unbundled Release: 3.0.1 Xref: Topic: Relevant Architectures: sparc NOTE: This patch does not address or correct issues pertaining to a standard SVR4 package. Therefore: 1) This patch is NOT installed via /usr/sbin/patchadd nor removed via /usr/sbin/patchrm. 2) This patch will NOT be reflected in the Solaris patch database or in /var/sadm/patch. 3) This patch will NOT be searchable or traceable via /usr/bin/showrev. Complete installation instructions are found below. BugId's fixed with this patch: Changes incorporated in this version: Patches accumulated and obsoleted by this patch: Patches which conflict with this patch: Patches required with this patch: Obsoleted by: Files included with this patch: README.121069-01 SP_4_0_0_1.bin Problem Description: Issues corrected in this patch -- 400001 Default number of streamer threads is too small Problem: In certain circumstances, such as when getting element details on a large number of hosts running multipathing software, there is no streamer thread available to return data to the client (JBoss). Symptom: Get Element Details requests may hang (deadlock) until the RMI timeout occurs. Solution: The default number of streamer threads has been changed from 10 to 20. This will be more than enough to cover our needs, since our default configuration specifies that no more than 10 synchronizer threads will run on the client for each CIMOM. 400002 Sun license must not enable provisioning by default Problem: Our contract with all OEMs does not give them any form of provisioning in the base license. Symptom: Wizard-based provisioning and provisiong from the CLI are available without setting additional license parameters. Solution: The code that checks if provisioning is enabled will now look at the path provisioning flag in the Sun license. 400003 Wrong storage pool totals for Symmetrix when there are metavolums Problem: For Symmetrix storage systems, capacity of any pool that contains meta-volumes is shown incorrectly. Symptom: The capacity shown is greater then the actual capacity by the amount equal to the sum of sizes of all the meta members (but not meta heads) in the pool Solution: The solution is to exclude meta-members when computing the size of a pool (meta-heads are still included in the calculation.) 400004 Capacity calculations fail after changing the "admin" passwd Problem: Default admin username and password were used for the CapacitySessionBean. When the admin password is changed, the bean no longer works, getting a security exception. Symptom: Clicking on Asset Management on the application screen throws an exception, 'javax.security.auth.login.LoginException: Unknown user: admin'. Solution: Use local connection for the bean which does not need username and password. 400005 RMITimeout is too small Problem: product shipped with an RMI Timeout that is too small for a medium or large-sized SAN Symptom: A Get Elements Detail request may time out when requesting large amounts of information. This can happen in a large SAN or with very large devices. Because of a new behavior in R4.0, once a request times out the remainder of a Get Elements Detail request is cancelled. Solution: Timeout period was changed to 3 hours. This is controllable by the system property RMITimeout. 400006 SWAPI provisioning guardrails Problem: zone set was activated on the wrong fabric Symptom: zone set was activated on the wrong fabric Solution: The code solution involves ensuring that the SWAPI handle used for provisioning refers to the correct fabric before the command is issued to activate the zone set. 400007 Cisco performance statistics are wrong Problem: The Cisco SNMP Provider uses a 32-bit counter to get port received and transmitted statistics. This is fine for real-time monitoring. However, when the port statistics collectors are running without real-time monitoring, the 32-bit counters can roll over multiple times between data collections. This causes the data to be meaningless. Symptom: The historical performance statistics charts will have meaningless data Solution: The solution is to get statistics data from the high capacity 64-bit counters. 400008 System Explorer tabs don't populate data Problem: The tree populating problems in System Explorer. Symptom: Sometimes the trees in List, Access and Path tabbed pan of System Explorer do not get populated completely. Solution: To fix this, we have increased the waiting period in tree populating thread from 50 milliseconds to 100 milliseconds. This allows more time for the Swing thread to create the controls before we attempt to populate them. 400009 HDS retries 5 times 20 minutes Problem: During discovery, there were no results after over an hour of waiting. The HiCommand server responded enough to allow us to recognize its existence and try to initiate communication, but then did not respond to HTTP requests. Symptom: Discovery does not complete in a timely manner. Solution: Reduce the retry timeout from 5 to 2. This will ensure that we don?t wait any longer than 40 mins when a request is made to a hung or unresponsive HiCommand server. 400010 Path snapshot is incomplete; Host Use of SAN storage reports are empty Problem: Path snapshot may be incomplete when there is large number of volume manager volumes and multipathing devices. Symptom: A number of reports are missing data. Examples include the Summary and Detail of the Host Use of SAN Storage Solution: Path snapshot records creation was changed to per path instead of all paths in one call. Path calculation for volume manager volumes and multipathing devices was optimized 400011 Materialized view refresh is too slow; LUNs per FA report has incorrect data Problem: The materialized views used by reports take a long time to populate if there are many storage volumes discovered. The MV_LUNSPERFASUMMARYVW view takes 80% of the total time. Symptom: Securepath Multipathing information doesn?t come up in the UI. The LUNs per FA report had incorrect data. Solution: Corrected and tuned the ?MV_LUNSPERFASUMMARYVW? view. Created missing index and tuned the view query. 400012 Unsupported version NaviCLI does not cause Clariion provider to disable itself Problem: Old versions of NaviCLI return incorrect information for MetaLUNs. The oldest version that AppIQ knows to work correctly is 6.5. When Storage Authority starts up, CLARiiON provider should check NaviCLI version, and disable itself if the version is older then 6.5. This defect resolution makes CLARiiON provider perform this check. Symptom: CLARiiON provider will work with NaviCLI versions older then 6.5. Solution: When CLARiiON provider starts up it will check the following system properties that can be configured in jboss.properties: cimom.provider.clariion.enableMetaLunSupport (default = true) cimom.clariion.minimumCliVersion.major (default = 6) cimom.clariion.minimumCliVersion.minor (default = 5) 400013 We don't recognize HDS WMS arrays Problem: Support for the Hitachi storage array family WMS was put in very late into the 4.0 release. Two minor items were missing: the wrong icon showed for the array in System Explorer and there were no HostMode2 options available when creating an HSG. Symptom: Wrong icon and no HostMode2 options when creating an HSG. Solution: Add checks for WMS storage array model in the UI code. 400014 Windows Proxy configuration page does not appear if no applications are licensed Problem: When the no applications are licensed, the Windows proxy tab (under Discovery -> Setup) is not visible even if the management server is running on a Solaris host. Symptom: The Windows proxy tab does not appear until MALS are purchased. Solution: The Windows Proxy tab no longer depends on the user having access to the Application Explorer. It will be visible whenever the management server is running on a non-Windows platform. (Solaris is the only supported non-Windows platform.) 400015, 400019 Missing indices on rollup tables, too many rollup records lead to slow purge after sync errors Problem: Rollup Tables are missing indices. Excess rollup records are created. Symptom: In certain error situations, the purge process of Get Element Details will be very slow, potentially taking hours to complete Solution: Created a primary key and relevant indices on rollup table columns. Added an additional key to the gcrollupconfig table. We now use separate rows in this table for garbage collection and rollup. 400016 System Explorer doesn't draw properly if a McData ISL does not include the destination port Problem: The McData SWAPI provider sometimes reports the node that a port is connected to but not the port it?s connected to. This can happen if a switch has an ISL to another switch that is not managed by the discovered EFCM. When this happened, System Explorer was unable to draw the physical topology, even for unaffected switches. Symptom: No physical topology would appear in System Explorer. The applet console would report a NullPointerException.. Solution: If a link is detected that has a destination node WWN but no destination port WWN, we will either 1.Draw a 'cloud port,' if the destination node was discovered; or 2.Ignore the link, if the destination node was not discovered. 400018 "Total Used" and "Available" capacity information are reversed Problem: Used and Free capacity of database application capacity summary in Capacity Explorer is reversed. Symptom: Reversed number for used and free capacity of database application summary. Solution: Corrected method call to calculate the summary data. Patch Installation Instructions: ----------------------------- 1) Execute /etc/init.d/appstormanager stop 2) Navigate to the directory where you unzipped this patch, and execute 120169-01/SP_4_0_0_1.bin 3) Follow the resulting on-screen patch installation instructions. 4) Execute /etc/init.d/appstormanager start Special Installation Notes: Either the "-i silent" or "-i console" switch may be used if needed. Version Notes: Navigate to the directory where the Sun StorEdge Operations Manager is installed. The file Version.txt contains the version string for the product. Base version string: 4.0.0.278 After patching: SP_4.0.0_n (where "n" is the revision number of the patch) Special Install Instructions: ------------------------------------ None. README -- Last modified date: Friday, October 21, 2005