Product SiteDocumentation Site

1.36. device-mapper-multipath

1.36.1. RHBA-2009:1645: bug fix update

Note

This update has already been released (prior to the GA of this release) as errata RHBA-2009:1645
Updated device-mapper-multipath packages that fix two bugs are now available.
The device-mapper-multipath packages provide tools to manage multipath devices by giving the device-mapper multipath kernel module instructions on what to do, as well as by managing the creation and removal of partitions for device-mapper devices.
This update addresses the following bugs:
* the udev rules for device-mapper-multipath were causing device-mapper to occasionally create multipath devices without using the user specified uid, gid, or mode. They have been replaced with equivalent rules that do not cause this issue. (BZ#537761)
* when LUNs were unmapped from LSI storage arrays, the multipath rdac path checker was not marking the paths as failed. This caused IO to the device to hang instead of fail. The rdac path checker now marks unmapped LUNs as failed. (BZ#538463)
Users are advised to upgrade to these updated device-mapper-multipath packages, which resolve these issues.

1.36.2. RHBA-2010:0255: bug fix and enhancement update

Updated device-mapper-multipath packages that fix several bugs and add various enhancements are now available.
The device-mapper-multipath packages provide tools to manage multipath devices using the device-mapper multipath kernel module.
This update applies the following bug fixes:
* The kpartx utility creates device maps from partition tables. Device-mapper devices with minor numbers greater than 255 caused kpartx to use the UUID from the wrong device when trying to create partitions. If the device had pre-existing partitions, kpartx would fail to create the new partitions. With this update, kpartx is now able to handle device-mapper devices with minor numbers greater than 255. (BZ#526550)
* The udev rules for device-mapper-multipath were causing device-mapper to occasionally create multipath devices without using the user specified uid, gid, or mode. They have been replaced with equivalent rules that do not cause this issue. (BZ#518575)
* When LUNs were unmapped from LSI storage arrays, the multipath rdac path checker was not marking the paths as failed. This caused IO to the device to hang instead of fail. The rdac path checker now marks unmapped LUNs as failed. (BZ#531744)
* The failover path grouping policy was not ordering the paths by priority, causing multipath to failover to the wrong path for devices with manual failback. The multipath paths are now correctly ordered with the failover path grouping policy. (BZ#537977)
* On some storage devices, if a LUN is deleted from an existing multipathd device, and a new LUN is presented to the host, it may end up with the same LUN ID and name as the old LUN. In this case, multipath will assume that this is the old LUN and belongs to the existing multipath device. This cause cause corruption. A new path checker "hp_tur" has been added that verifies the WWID of the LUN when it checks the path, to avoid this problem. (BZ#437585)
* The "tur" path checker was marking paths in standby mode as "failed". It now correctly marks them as "ghost". (BZ#473039)
* Multipath wasn't correctly showing device renames in dry-run mode. This has been fixed. (BZ#501019)
* Multipath was incorrectly setting the hardware handler for HP StorageWorks devices.This has been fixed. (BZ#475967)
* On some storage devices, multipath would display incorrect path information the first time multipath listed the paths after recovery. This has been fixed. (BZ#499080)
* If a path is removed while it is still part of a multipath device, it was taking multipath minutes to mark it as failed. This should now happen immediately at the end of the next path checking interval. (BZ#527754)
* the multipathd daemon needs constant access to /var/lib and /var/run. However it was not allowing any devices mounted under /var to be removed. Now it only keeps open what it needs. (BZ#532424)
* the multipath checker functions were not using the default scsi timeouts. Instead, each checker set its own timeout. Now all checker functions with explicit timeouts use the scsi timeout set it /sys/block/sd<x>/device/timeout by default. This can be changed by setting the "checker_timeout" option in /etc/multipath.conf(BZ#553042)
* Multipathd was printing extraneous error messages. This has been fixed. (BZ#472171, BZ#502128, BZ#524178)
* The multipath man page had some mistakes and missing information. This has been fixed. (BZ#481239, BZ#510331, BZ#554830)
* A locking error could cause multipathd to deadlock if it failed to create a multipath device correctly. This has been fixed (BZ #537281)
This update adds the following enhancements:
* Default configurations were added for more IBM, HP, SUN, and DELL devices. (BZ#504619i, BZ#512243, BZ#515171, BZ#517896, BZ#540882,
* The kpartx utility now supports DASDs devices with more then 65520 cylinders. (BZ#524009)
All users are advised to upgrade to these updated packages, which resolve these issues and add these enhancements.

Note: This documentation is provided {and copyrighted} by Red Hat®, Inc. and is released via the Open Publication License. The copyright holder has added the further requirement that Distribution of substantively modified versions of this document is prohibited without the explicit permission of the copyright holder. The CentOS project redistributes these original works (in their unmodified form) as a reference for CentOS-5 because CentOS-5 is built from publicly available, open source SRPMS. The documentation is unmodified to be compliant with upstream distribution policy. Neither CentOS-5 nor the CentOS Project are in any way affiliated with or sponsored by Red Hat®, Inc.