Patch-ID# 106707-05 Keywords: A1000 A3000 A3500 cluster Synopsis: RAID Manager 6.1.1: A1000/A3x00 Support for Sun Cluster 2.x Date: Jun/05/2001 Solaris Release: 2.5.1 2.6 SunOS Release: 5.5.1 5.6 Unbundled Product: RAID Manager Unbundled Release: 6.1.1 NOTE: update 1 and update 2 Xref: Topic: Relevant Architectures: sparc BugId's fixed with this patch: 4152054 4168155 4168934 4169133 4246748 4299223 4378017 4400814 4449065 Changes incorporated in this version: 4449065 Patches accumulated and obsoleted by this patch: Patches which conflict with this patch: 106552-01 108555-01 Patches required with this patch: 106513-03 or greater Obsoleted by: Files included with this patch: /etc/osa/rmparams /kernel/drv/rdriver /sbin/rdaemon /usr/share/man/man4/rdriver.4 /usr/share/man/man4/rmparams.4 Problem Description: 4449065 README.106707-04 doesn't indicate sun cluster setting for Rdac_NoAltOffline (from 106707-04) 4378017 106707-03 removes the only copy of the .bwd on the entire system 4400814 Patch 106707 re-installs the file /etc/osa/rmparams (from 106707-03) 4299223 106707 Synopsis is incomplete, not mentioning SC 2.2 (from 106707-02) 4246748 RM 6.1.1 needs tx chip firmware, since zx chip shortage (from 106707-01) 4169133 "fsck -F vxfs ... failed" during haswitch if luns were moved w/o "vxdctl enable " 4152054 Disk pull- RAID 0 Lun of A1000 causes the node to abort due to interconnect fai l 4168155 Rdac driver misbehavior when powering off a mirrored sonoma in a SC config. 4168934 Losing power to A3000 will cause rdac and sd to go into a weird states Patch Installation Instructions: -------------------------------- Refer to the Install.info file for instructions on using the generic 'installpatch' and 'backoutpatch' scripts provided with each patch. Any other special or non-generic installation instructions should be described below as special instructions. Special Install Instructions: ----------------------------- Patch 106513-03 must be installed prior to installation of this patch because it is the RM 6.1.1 jumbo patch. Starting with 106513-03 there is firmware included therein that must be downloaded to make this patch operative. This patch provides a new rmparams file. The user should copy his /etc/osa/rmparams file to a safe place before installing this patch. He should then merge his previous changes into the new rmparams file. Once this patch has been installed, the user must do a system boot (-r argument not required) to insure the new drivers and daemons are loaded. The Rdac_RetryCount parameter in rmparams will be set to 1 after this patch has been applied in order to prevent cluster node removal when RDAC fails a controller. Note: the default value is 7 The Rdac_NoAltOffline parameter in rmparams will be set to TRUE after this patch has been applied in order to prevent cluster node removal when RDAC fails a controller. Note: the default value is FALSE New RDAC rmparams parameters Rdac_RetryCount: Specifies the number of times the RDAC driver/daemon will retry an errorred I/O before giving up and returning it to the system. The default is 7. (For Solaris cluster environments: in order to prevent cluster node removal when RDAC fails a controller. Rdac_RetryCount will be set to 1 after this patch has been applied.) Rdac_NoAltOffline: If TRUE, the RDAC daemon will mark a LUN as offline when it gives up retrying an I/O and the alternate path is not available. This will result in all I/Os to that LUN being errorred immediately without retries. The default is FALSE which means every I/O will be retried Rdac_RetryCount times. (For Solaris cluster environments: in order to prevent cluster node removal when RDAC fails a controller. Rdac_NoAltOffline will be set to TRUE after this patch has been applied.) Rdac_Fail_Flag This parameter is not new, but will now be used by RDAC in the failover routine. If the flag is FALSE, controllers will not be failed, but just made passive. README -- Last modified date: Tuesday, June 5, 2001