* * * READ ME * * * * * * Veritas Low Latency Transport 6.2.1 * * * * * * Patch 600 * * * Patch Date: 2016-11-29 This document provides the following information: * PATCH NAME * OPERATING SYSTEMS SUPPORTED BY THE PATCH * PACKAGES AFFECTED BY THE PATCH * BASE PRODUCT VERSIONS FOR THE PATCH * SUMMARY OF INCIDENTS FIXED BY THE PATCH * DETAILS OF INCIDENTS FIXED BY THE PATCH * INSTALLATION PRE-REQUISITES * INSTALLING THE PATCH * REMOVING THE PATCH PATCH NAME ---------- Veritas Low Latency Transport 6.2.1 Patch 600 OPERATING SYSTEMS SUPPORTED BY THE PATCH ---------------------------------------- RHEL6 x86-64 PACKAGES AFFECTED BY THE PATCH ------------------------------ VRTSllt BASE PRODUCT VERSIONS FOR THE PATCH ----------------------------------- * Symantec Cluster Server 6.2 * Symantec Storage Foundation Cluster File System HA 6.2 * Symantec Storage Foundation for Oracle RAC 6.2 * Symantec Storage Foundation HA 6.2 SUMMARY OF INCIDENTS FIXED BY THE PATCH --------------------------------------- Patch ID: 6.2.1.600 * 3905431 (3905430) Application IO hangs in case of FSS with LLT over RDMA during heavy data transfer. Patch ID: VRTSllt-6.2.1.500 * 3877463 (3877459) Veritas Cluster Server (VCS) does not support Red Hat Enterprise Linux 6 Update 8 (RHEL6.8). Patch ID: VRTSllt-6.2.1.100 * 3794201 (3794154) Veritas Cluster Server (VCS) does not support Red Hat Enterprise Linux 6 Update 7 (RHEL6.7). Patch ID: VRTSllt-6.2.1.000 * 3663730 (3567354) The Linux kernel panics when it receives a shared non-linear socket buffer (skb) for linearization from the Low Latency Transport (LLT). * 3663741 (3663740) In a rare scenario, divide by zero error is displayed when lltshow -p command is run. DETAILS OF INCIDENTS FIXED BY THE PATCH --------------------------------------- This patch fixes the following incidents: Patch ID: 6.2.1.600 * 3905431 (Tracking ID: 3905430) SYMPTOM: Application IO hangs in case of FSS with LLT over RDMA during heavy data transfer. DESCRIPTION: In case of FSS using LLT over RDMA, sometimes IO may hang because of race conditions in LLT code. RESOLUTION: LLT module is modified to fix the race conditions arising due to heavy load with multiple application threads. Patch ID: VRTSllt-6.2.1.500 * 3877463 (Tracking ID: 3877459) SYMPTOM: Veritas Cluster Server (VCS) does not support Red Hat Enterprise Linux 6 Update 8 (RHEL6.8). DESCRIPTION: VCS did not support RHEL versions released after RHEL6 Update 7. RESOLUTION: VCS support for Red Hat Enterprise Linux 6 Update 8 (RHEL6.8) is now introduced. Patch ID: VRTSllt-6.2.1.100 * 3794201 (Tracking ID: 3794154) SYMPTOM: Veritas Cluster Server (VCS) does not support Red Hat Enterprise Linux 6 Update 7 (RHEL6.7). DESCRIPTION: VCS did not support RHEL versions released after RHEL6 Update 6. RESOLUTION: VCS support for Red Hat Enterprise Linux 6 Update 7 (RHEL6.7) is now introduced. Patch ID: VRTSllt-6.2.1.000 * 3663730 (Tracking ID: 3567354) SYMPTOM: Low Latency Transport (LLT) observes a kernel panic on receiving a shared non-linear skb. DESCRIPTION: When LLT receives a shared non-linear skb from the network driver, it passes the skb to the kernel for linearization. Since the skb is shared between multiple owners, the kernel cannot linearize it, which causes a system panic. RESOLUTION: The code is modified such that on receiving a non-linear skb, LLT validates if the skb is shared. If the validation is true, then the kernel clones the skb, returns the cloned skb to LLT, and decreases the reference count of the original skb. Further, LLT passes the cloned skb to the kernel for linearization. * 3663741 (Tracking ID: 3663740) SYMPTOM: In a rare scenario, the system shows a divide by zero error and generates a core dump when the lltshow -p command is run. DESCRIPTION: If a variable value becomes zero under certain conditions and lltshow p is run at the same time to get the port statistics, the system shows a divide by zero error. RESOLUTION: The code is modified to skip printing the batch upcall latency average when the batch count is zero. INSTALLING THE PATCH -------------------- Run the Installer script to automatically install the patch: ----------------------------------------------------------- Please be noted that the installation of this P-Patch will cause downtime. To install the patch perform the following steps on at least one node in the cluster: 1. Copy the patch llt-rhel6_x86_64-Patch-6.2.1.600.tar.gz to /tmp 2. Untar llt-rhel6_x86_64-Patch-6.2.1.600.tar.gz to /tmp/hf # mkdir /tmp/hf # cd /tmp/hf # gunzip /tmp/llt-rhel6_x86_64-Patch-6.2.1.600.tar.gz # tar xf /tmp/llt-rhel6_x86_64-Patch-6.2.1.600.tar 3. Install the hotfix(Please be noted that the installation of this P-Patch will cause downtime.) # pwd /tmp/hf # ./installVRTSllt621P6 [ ...] You can also install this patch together with 6.2.1 maintenance release using Install Bundles 1. Download this patch and extract it to a directory 2. Change to the Veritas InfoScale 6.2.1 directory and invoke the installmr script with -patch_path option where -patch_path should point to the patch directory # ./installmr -patch_path [] [ ...] Install the patch manually: -------------------------- Perform the following steps on the node on which this patch is being applied: 1. Take backup of your configurations 2. Stop VCS 3. Install the patch 4. Restart VCS on the node Stopping VCS on the cluster node -------------------------------- 1. Ensure that the /opt/VRTSvcs/bin directory is included in your PATH environment variable so that you can execute all the VCS commands. Refer to Veritas Cluster Server Installation Guide for more information. 2. Freeze all the service groups persistently. # haconf -makerw # hagrp -freeze [group] -persistent 3. If the cluster is writable, you may close the configuration before stopping the cluster. Stop the cluster on this node: # haconf -dump -makero run the following command: # hastop -local -force Verify that the cluster is stopped by running the ha command: # hasys -state make sure that both had and hashadow processes are stopped. Also, stop the VCS CmdServer: # CmdServer -stop 4. Stop the cluster services in order: - For VxFEN: # /etc/init.d/vxfen stop - For GAB: # /etc/init.d/gab stop - For LLT # /etc/init.d/llt stop Note: Stop any other services depending on GAB Installing the Patch -------------------- 1. Un-compress the downloaded patch and change the directory to the uncompressed patch location. Install the patch using the following command: # rpm -Uvh VRTSllt-6.2.1.500-RHEL6.x86_64.rpm 2. Run the following command to verify if the new patch has been installed: # rpm -q VRTSllt If the proper patch is installed, the following output is displayed: VRTSllt-6.2.1.500-RHEL6.x86_64 Re-starting the cluster services: -------------------------------- 1. To start the cluster services in order run the following command: - For LLT: # /etc/init.d/llt start - For GAB: # /etc/init.d/gab start - For VxFEN: # /etc/init.d/vxfen start - For VCS: # hastart # /opt/VRTSvcs/bin/CmdServer 2. Make VCS cluster writable: # haconf -makerw 3. Unfreeze all the groups: # hagrp -unfreeze [group] -persistent # haconf -dump -makero REMOVING THE PATCH ------------------ Removal of the patch will result in removing the 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. Run the following steps on the node where the patch is being uninstalled: To remove the patch from a cluster node: --------------------------------------------- 1. Freeze all the service groups persistently. # haconf -makerw # hagrp -freeze [group] -persistent 2. Stop VCS by following the stop sequence as mentioned in the PATCH INSTALLATION section. 3. Remove the patch by using the following command: # rpm -e VRTSllt --nodeps 4. Verify that the patch has been removed from the system: # rpm -qa|grep VRTSllt Ensure that the package is not displayed. This confirms that the package is removed. 5. Install the relevant package from the VCS base media. Also refer to SORT for any applicable patches. 6. To start the cluster service, follow the start sequence as mentioned in the PATCH INSTALLATION section. 7. Unfreeze all the groups. # hagrp -unfreeze [group] -persistent # haconf -dump -makero SPECIAL INSTRUCTIONS -------------------- NONE OTHERS ------ NONE