User manual HP STORAGEWORKS SECURE PATH FOR SUN SOLARIS SOFTWARE STORAGEWORKS SECURE PATH V3.0B FOR SUN SOLARIS RELEASE NOTES (AA-RKYEL-TE, APRI

[. . . ] release notes hp StorageWorks secure path version 3. 0b for Sun Solaris Product Version: 3. 0B Product Version: Tenth Edition (April 2003) Part Number: AA-RKYEL-TE This document summarizes features and characteristics of hp StorageWorks Secure Path Version 3. 0B for Sun Solaris systems, using StorageWorks Array Controllers for Fibre Channel switched fabric. For the latest version of these Release Notes and other Secure Path documentation, access the HP storage website at: http://www. hp. com/country/us/eng/prodserv/storage. html. © Hewlett-Packard Company, 2003. Hewlett-Packard Company makes no warranty of any kind with regard to this material, including, but not limited to, the implied warranties of merchantability and fitness for a particular purpose. Hewlett-Packard shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. [. . . ] This results in being unable to reestablish communication with the array following reconnection of the array. The timer can be restarted by quiescing any known array object (adapter, controller or path) using spmgr quiesce followed by spmgr restart all. 14 Secure Path Version 3. 0B for Sun Solaris Release Notes Avoiding Problem Situations Removing an HSG80 disk If a configured disk is physically removed from an HSG80 based array, a specific SCSI Unit Attention is broadcast to all LUNs on all paths. Secure Path is mistakenly interpreting this as a path failure event such that all paths to all LUNs are momentarily failed. The last gasp logic immediately corrects the condition on all good LUNs and I/O is only blocked for less than a second. Secure Path's event notification logs all these events. To prevent this situation, be sure to delete any device from the array before removing it. Assigning Grouped LUNs to Different Servers Although LUNs that are members of a partitioned group may be assigned to different servers, it is recommended that grouped LUNs under Secure Path control be assigned to the same server. Multiple path failures on one server can cause a controller failover that is not recognized by the other server. In extreme cases, both servers may compete for different controllers, resulting in ping-pong failovers between controllers. ACS 8. 7S parent WWLUNID The ACS 8. 7S feature that allows creation of a snapshot or clone using it's parent WWLUNID (two devices with the same WWLUNID) is not supported by Secure Path. Attempting to manage these devices with spmgr results in potential confusion between the units. Clustering with Solaris 2. 6 When using Sun clusters with Solaris 2. 6, you may get the following error message when you attempt to run spmgr commands: Error: This host is not authorized to connect to remote host. This is due to the fact that spagent is using the actual server hostname for authentication, and not the cluster host name. To fix this problem, add the cluster name as an authorized client by using the spmgr client add <cluster_name> command. Dynamic Reconfiguration with Solaris 2. 6 Dynamic Reconfiguration (DR) is not supported with Secure Path Version 3. 0B on servers running Solaris 2. 6. Secure Path Version 3. 0B for Sun Solaris Release Notes 15 Avoiding Problem Situations JNI link up failure after portDisable If Fibre Channel switch utilities are used to test JNI HBA failover configurations, enabling a switch port (portEnable), after disabling the port (portDisable) may fail to reestablish the link with the JNI HBA. If this occurs, quick-boot the switch (fastboot) to clear the problem. Keep Path Verification Enabled If Path Verification is disabled and the last available path fails, the last gasp feature may timeout before completing repair of all good paths. To avoid this situation, HP recommends that Path Verification be enabled (the default setting) with a verification period of not greater than 90 seconds. Spmgr display blocks I/O if failed path on unattached LUN If units are in the unattached list (spmgr display ­u) and one or more paths to the unattached units has failed, spmgr display hangs and I/O is blocked for approximately 30 seconds per unattached failed path. I/O is only blocked on the array with the unattached failed path. If this is a concern, unpresent all LUNs after deleting them using spmgr delete. The units will then not be shown in the unattached list. Spagent Core File After Upgrade Installation During the Version 3. 0B upgrade installation, the install process occasionally causes the existing spagent daemon to crash and generate a core file in the root directory. The upgrade install completes normally and the subsequent reboot starts the new spagent correctly. You can safely remove the resultant core file, which is less than 1 MB. Spmgr select and restore changes in Version 3. 0B Spmgr commands select and restore operate differently in Version 3. 0B than they did in previous versions. Additionally, the spmgr prefer and spmgr unprefer commands have been removed and replaced by having select assume the prefer functionality. Using spmgr select and spmgr restore on LUNs that are part of a partitioned HSG80 storageset results in all LUNs of that partition being selected or restored. If the path being selected is on the opposite controller from the currently active path, the spmgr select ­f -c controller_serial_number operation (with the ­f option to force the controller change) causes the HSG80 to move control of the storageset to that 16 Secure Path Version 3. 0B for Sun Solaris Release Notes Avoiding Problem Situations controller. All LUNs that are partitions of that storageset will also be moved. [. . . ] Run /opt/steam/bin/config. sh to modify the old SWCC access device to the new access device. Enter the position number of the subsystem you wish to modify. Enter the new access device in the form of cXtYdZs2. Enter D to exit if your information is correct. [. . . ]