OS: Linux OS Version: RHEL50 (i686/x86_64) Fixes Applied for Products: VRTSvcs - Veritas Cluster Server by Symantec VRTSvcsag - Veritas Cluster Server Bundled Agents by Symantec Additional Instructions: Please read the instructions below before installing the patch. PATCH VRTSvcs and VRTSvcsag 5.0MP3RP2HF5 for VCS 5.0MP3RP2 =============================================================== Patch Date: September, 2009 This README provides information on: * BEFORE GETTING STARTED * CRC AND BYTE COUNT * FIXES AND ENHANCEMENTS INCLUDED IN THE PATCH * PACKAGES AFFECTED BY THE PATCH * INSTALLING THE PATCH * UNINSTALLING THE PATCH BEFORE GETTING STARTED: ---------------------- This patch applies only to: VRTSvcsag running on RHEL50 (i686/x86_64) VRTSvcs running on RHEL50 (i686/x86_64) Ensure that you are running the supported configurations before installing this patch. CRC AND BYTE COUNT: ------------------ Ensure that the patch file you have downloaded matches the following checksum and byte count. The following command can be used to find the checksum: # cksum VRTSvcs-5.0.30.25-MP3RP2HF5_RHEL5.i686.rpm 2978064712 31665242 VRTSvcs-5.0.30.25-MP3RP2HF5_RHEL5.i686.rpm # cksum VRTSvcsag-5.0.30.25-MP3RP2HF5_RHEL5.i686.rpm 2257238690 491377 VRTSvcsag-5.0.30.25-MP3RP2HF5_RHEL5.i686.rpm FIXES AND ENHANCEMENTS INCLUDED IN THE PATCH: -------------------------------------------- Etrack Incidents: 1824996, 1789008, 1730192, 1723347, 1557126, 1638428, 1665814, 1676451 SDR's of Fixed Symantec Incidents: -------------------------------- Symantec Incident : 1824996 Symptom: On a loaded system, it was found that, CFSfsckd agent deadlocked in ptmalloc_lock_all() function. Defect Description: On a loaded system, the CFSfsckd agent was not able to heartbeat with engine and after waiting for "AgentReplyTimeout" value engine sends the SIGABRT signal to agent. When the agent receives the SIGABRT signal, the agent invokes the signal handler and the signal handler calls fork(2) system call. Linux OS does not support fork(2) in signal handlers, due to which deadlock occurred. Resolution: Removed the fork(2) call from the signal handler. Symantec Incident : 1789008 Symptom: Presence of file /var/lib/nfs/etab, upon panic reboot causes incorrect operation of Share agent. Defect Description: When the system panics, the /var/lib/nfs/etab file is not cleared at boot-time. This results in VCS incorrectly seeing share resources as online. The Share agent service group cannot be failed over to the other node as the service group goes into partial state. The NFS client receives an “Stale NFS File Handle” error. Resolution: A new mandatory attribute is added to Share agent type definition named “NFSRes”. This attribute is the NFS Resource configured on the system. This ensures that the /var/lib/nfs/etab file is cleared during the First Probe of Share resource, if NFS resource is not yet up. Description of NFSRes attribute: NFSRes, the NFS resource be configured on the system. This attribute is required to determine the current state of the NFS Resource. Type and dimension: string-scalar Sample Configuration: NFS nfs ( Address = "[ip-address]" ) Share share ( NFSRes = nfs PathName = "/nfs_mnt" Options = "rw" ) Previous Patch Fixes: Symantec Incident: 1730192 Symptom: DNS clean never used to succeed. Defect Description: The DNS clean entry point was not parsing the argument list correctly. Hence, it always reported the clean operation as failed. VCS kept on re-attempting clean operation and this sequence never used to end. Resolution: Fixed the clean entry point code to parse the argument list correctly. Now the DNS resource's clean operation should succeed, if there is no fault. Symantec Incident: 1723347 Symptom: Mount agent returns offline even when the resource is mounted. Defect Description: Mount agent did not support large file systems. Thus statvfs(3P) call within large file systems used to fail. Due to this, the monitor entry point used to return offline status. Resolution: Fixed the Mount agent to support large file systems and return correct status. Symantec Incident: 1557126 The VCS Application agent was not able to recognize a running process containing special characters like round brackets. Symantec Incident: 1638428 Mount agent doesn't support NFS version 4 in Linux, even though NFS agent supports it. Symantec Incident: 1665814 The offline entry point of NFSRestart does not stop NFS daemons. Symantec Incident: 1676451 NFS clients were getting "permission denied" errors after NFS service group failover. PACKAGES AFFECTED BY THE PATCH: ------------------------------- This patch updates the following VCS package(s) VCS from 5.0MP3RP2(5.0.3.2) or later to 5.0MP3RP2HF5(5.0.3.25) INSTALLING THE PATCH: -------------------- The following steps should be run on all nodes in the VCS cluster: Stopping VCS on the cluster node: -------------------------------- 1. Ensure that "/opt/VRTS/bin" directory is included in your PATH environment variable so you can execute all the VCS commands. Refer to Veritas Cluster Server Installation Guide for more information. 2. Verify that the version of VRTSvcs and VRTSvcsag is 5.0MP3RP2 or later for Linux. 3. Freeze all the service groups persistently. # haconf -makerw # hagrp -freeze [group] -persistent # haconf -dump -makero 4. Stop the cluster on all nodes. If the cluster is writable, you may close the configuration before stopping the cluster. # haconf -dump -makero From any node, execute the following command. # hastop -all or # hastop -all -force Verify that the cluster is stopped on all nodes by running the ha command. # hasys -state On all nodes, make sure that both had and hashadow processes are stopped. Also, stop the VCS CmdServer on all nodes. # CmdServer -stop 5. Move /etc/VRTSvcs/conf/config/types.cf file to /etc/VRTSvcs/conf/config/types.cf.orig. Installing the Patch: -------------------- 1. Un-compress the downloaded patch from Symantec. Change directory to the unzipped patch location. Install the VRTSvcs and VRTSvcsag patch using the following command: # rpm -Uvh VRTSvcs-5.0.30.25-MP3RP2HF5_RHEL5.i686.rpm # rpm -Uvh VRTSvcsag-5.0.30.25-MP3RP2HF5_RHEL5.i686.rpm 2. Verify that the new patch has been installed: # rpm -q VRTSvcs # rpm -q VRTSvcsag You will find the following output on display with the patch installed properly: VRTSvcs-5.0.30.25-MP3RP2HF5 VRTSvcsag-5.0.30.25-MP3RP2HF5 Re-starting VCS on the cluster node: ----------------------------------- Manual configuration update steps: 1. Adding NFSRes attribute to all the systems: Add NFSRes attribute to every Share resource configuration in main.cf. Sample Configuration, that need to be added : NFS nfs ( Address = "[ip-address]" ) Share share ( NFSRes = nfs PathName = "/nfs_mnt" Options = "rw" ) Verify consistency of main.cf file after updation: # hacf -verify /etc/VRTSvcs/conf/config/ Note: Please make sure that the NFSRes attribute is added to every Share resource that is configured on the system, after applying this patch. In case of attribute mismatch error for NFSRes attribute, please verify the types.cf is upgraded properly. 2. At this time, you can start the cluster services on all cluster nodes. First start VCS on one node: # hastart On all the other nodes, start VCS by issuing hastart after the first node goes to LOCAL_BUILD or RUNNING state. 3. Unfreeze all the groups. # haconf -makerw # hagrp -unfreeze [group] -persistent # haconf -dump -makero OR Runtime configuration update steps: 1. At this time, you can start the cluster services on all cluster nodes. First start VCS on one node: # hastart On all the other nodes, start VCS by issuing hastart after the first node goes to LOCAL_BUILD or RUNNING state. 2. Make VCS cluster Writable: # haconf -makerw 3. To modify the configuration at run-time, instead of modifying main.cf manually please execute the following commands: NOTE: Please make sure that the Service groups are freezed. # for res in `hatype -resources Share`; do hares -modify $res NFSRes [NFS_Resource_Name] done Note: Please make sure that the NFSRes attribute is added to every Share resource that is configured on the system, after applying this patch. In case of attribute mismatch error for NFSRes attribute, please verify the types.cf is upgraded properly. 4. Unfreeze all the groups. # hagrp -unfreeze [group] -persistent # haconf -dump -makero UNINSTALLING THE PATCH: ----------------------- Removal of the patch will result in removing whole package from the system/node. To go back to a previous installed version of the package, you may need to re-install the package again. Following steps need to be run on all the VCS cluster node: Steps to remove the Patch from a cluster node: --------------------------------------------- 1. Follow the steps 1-4 provided under "Stopping VCS on the cluster node" section above, to stop VCS on the node and unload any drivers, as required. 2. Remove the patch by the following command: # rpm -e VRTSvcs # rpm -e VRTSvcsag 3. Verify that the patch has been removed from the system: # rpm -qa | grep VRTSvcs # rpm -qa | grep VRTSvcsag The package VRTSvcsag should not be displayed, which confirms that the package is removed. 4. Install VRTSvcs and VRTSvcsag package from VCS 5.0MP3 Installer CD and upgrade to VCS 5.0MP3RP2. 5. At this time, you can start the cluster services on all cluster nodes. First start VCS on one node: # hastart On all the other nodes, start VCS by issuing hastart after the first node goes to LOCAL_BUILD or RUNNING state. 6. Unfreeze all the groups. # haconf -makerw # hagrp -unfreeze [group] -persistent # haconf -dump -makero