Introduction
This document describes how SAP HIgh Performance Analytic Appliance (HANA) scale-out solutions with Electromagnetic Compatability (EMC) VNX storage and applies to the Cisco SAP HANA appliance as well as to SAP HANA Tailored Datacenter Integration (TDI) implementations. It also explains how multipath configuration changes to mitigate unexpected outages of the environment.
Prerequisites
Requirements
There are no specific requirements for this document.
Components Used
This document is not restricted to specific software and hardware versions.
The information in this document was created from the devices in a specific lab environment. All of the devices used in this document started with a cleared (default) configuration. If your network is live, make sure that you understand the potential impact of any command.
Problem: Unexpected outages of the SAP HANA scale-out environment
For SAP HANA the required data and log file systems are mounted via fiber channel from the EMC VNX storage. Mounting and unmounting of the devices is controlled by the SAP HANA storage API and is based on the Linux Device Mapper Multipath (DM-MP) module. The configuration consists of eight paths (active/passive) to each storage device.
With SUSE Linux Enterprise System 11.3, Linux kernel 3.0.101-0.40 a bug in the multipathing logic has been detected where the medium access timeout counter is not being reset. Due to the bug the storage device will be taken offline when the counter reaches a value of 2.
This may result in unexpected outages of the SAP HANA scale-out environment.
The bug is not encountered as long as the hardware and cabling is error free.
Solution
Update the SUSE Linux Enterprise System 11.3 kernel to version 3.0.101-0.47.52.1 or above and adjust the following parameters in the /etc/multipath.conf file:
rr_min_io
flush_on_last_del
fast_io_fail_tmo
dev_loss_tmp
[an error occurred while processing this directive]
The resulting configuration file must look like this:
defaults {
user_friendly_names no
}
devices {
device {
vendor "DGC"
product ".*"
product_blacklist "LUNZ"
features "0"
hardware_handler "1 emc"
path_selector "round-robin 0"
path_grouping_policy group_by_prio
failback immediate
rr_weight uniform
no_path_retry 5
rr_min_io 1000
path_checker emc_clariion
prio emc
flush_on_last_del yes
fast_io_fail_tmo off
dev_loss_tmo 120
}
}
[an error occurred while processing this directive]
Note: The EMC white paper VNX Configuration Recommendations for SAP HANA TDI, page 20, does not mention these changes because the white paper had been created before the kernel bug was detected.
Related Information
Feedback