[El-errata] ELBA-2012-2017 Oracle Linux 5 device-mapper-multipath bug fix update

Errata Announcements for Oracle Linux el-errata at oss.oracle.com
Fri Jun 1 08:41:33 PDT 2012


Oracle Linux Bug Fix Advisory ELBA-2012-2017

The following updated rpms for Oracle Linux 5 have been uploaded to the 
Unbreakable Linux Network:

i386:
device-mapper-multipath-0.4.9-46.0.4.el5.i386.rpm
device-mapper-multipath-libs-0.4.9-46.0.4.el5.i386.rpm
kpartx-0.4.9-46.0.4.el5.i386.rpm

x86_64:
device-mapper-multipath-0.4.9-46.0.4.el5.x86_64.rpm
device-mapper-multipath-libs-0.4.9-46.0.4.el5.x86_64.rpm
kpartx-0.4.9-46.0.4.el5.x86_64.rpm

SRPMS:
http://oss.oracle.com/ol5/SRPMS-updates/device-mapper-multipath-0.4.9-46.0.4.el5.src.rpm


Description of changes:

[0.4.9-46.0.4.el6]
- Correct the check mask for PQ in rdac path checker (Chaskiel Grundman) 
[orabug 14061177]
- Service mode changes for rdac storage (Babu Moger) [orabug 14061177]
- Blacklist all the management Luns by default (Babu Moger) [orabug 
14061177]
- Lower the log level for rdac TAS messages (Babu Moger) [orabug 14061177]

[0.4.9-46.0.3.el6]
- Provide a copy of /sbin/multipath.static, as both /etc/rc.sysinit and
    /etc/init.t/netfs call it, which is no longer built by 0.4.9 (John 
Sobecki)
    [orabug 13817465]

[0.4.9-46.0.2.el6]
- add patch to revert mpatha to mpath0 numeric names (John Sobecki) 
[orabug 10282894]
- add patch to fix "invalid number of priority groups" error (Zhenzhong 
Duan) [orabug 11884664]
- fix page 83 options for EMC and other storage. (Guru Anbalagane) 
[orabug 10364063]
- revert all scsi_id callouts to el5 behavior (Guru Anbalagane) [orabug 
10364063]
- Add udev rule to create udev socket for multipathd (Xiaowei Hu) 
[orabug 10121864]
- Remove path_checker from multipath.conf (Xiaowei Hu) [orabug 10121963]
- fixed kpartx not consistently creating devices bug (Xiaowei Hu) 
[orabug 10250424]
- During upgrade, restore the .rpmsave copy of multipath.conf if no config
   file is present (John Sobecki) [orabug 10275766]
- port EL5u5 multipath rules (Chuck Anderson) [orabug 10105043]
- Add kpartx.static back - needed for kernel 2.6-32 (John Sobecki) 
[orabug 10104535]
- Add revert-scsi-id-checker-default.patch - scsi_id patch (John 
Sobecki) [orabug 10096746]
- convert the "!" in cciss device names to "/" (John Sobecki) [RHBZ 
729478] [orabug 13481064]

[0.4.9-46.0.1.el6]
- Add patch to remove path_checker from the multipaths section of the
   default multipath.conf.default (John Sobecki) [orabug 10650061]

[-0.4.9.46]
- Remove 0107-RHBZ-725541-01-async-tur-checker.patch
   * This patch caused a regression where, multpathd could corrupt is memory
     when a path was deleted (bz747604)
- Refresh 0110-RHBZ-725541-04-serialize-startup.patch
-Resolves: bz #747604

[-0.4.9.45]
- Add 0118-RHBZ-738298-revert-631009.patch
   * reverts most of 
0101-RHBZ-631009-disable-udev-disk-rules-on-reload.patch
     but still allows the -u option, for compatibility
- Resolves: bz #738298

[-0.4.9.44]
- Add 0116-RH-manpage-update.patch
   * update the manpages with new features
- Add 0117-RHBZ-697386-logger-shutdown-crash.patch
   * The main multipathd thread was racing with the logger, now it
     waits for the logger thread to shutdown first.
- Resolves: bz #697386

[-0.4.9.43]
- Add 0102-RHBZ-714821-dont-remove-map-twice.patch
   * multipathd was removing a map twice in one code path, causing
     double free errors
- Add 0103-RHBZ-725541-09-no-partitions.patch
   * if a device has the "no_partitions" feature, kpartx ignores it
- Add 0104-RHBZ-725541-10-force-devmap.patch
   * override "no_partitions"
- Add 0105-RHBZ-725541-08-reset-running.patch
   * abort timed-out IO in directio checker
- Add 0106-RHBZ-725541-07-tur-checker-reservation-conflict.patch
   * don't count a path as down if it has a reservation conflict. The
     kernel should be failing these errors up past multipath now.
- Add 0107-RHBZ-725541-01-async-tur-checker.patch
   * Make tur checker async
- Add 0108-RHBZ-725541-06-offline-check.patch
   * Don't get the priority of offline devies.
- Add 0109-RHBZ-725541-05-only-count-up-and-ghost-paths.patch
   * fix path counting
- Add 0110-RHBZ-725541-04-serialize-startup.patch
   * add the "show daemon" multipathd command, so users can check
     that the daemon has started up completely.
- Add 0111-RHBZ-725541-03-handle-LBA_DEPENDENT-state.patch
   * handle LBA_DEPENDENT alua state.
- Add 0112-RHBZ-725541-02-infinite-dev-loss-tmo.patch
   * allow dev_loss_tmo to be set to "inifinty", which sets it to
     the maximum allowed value.
- Add 0113-RHBZ-719571-validate-guid-partitions.patch
   * don't build partition devices for invalid gpt's.
- Add 0114-RHBZ-723168-adjust-messages.patch
- Add 0115-RHBZ-713754-lower-minio.patch
   * add rr_min_io_rq parmeter, that sets the min_io to 1 for
     request based multipath
Resolves: bz #699577, #713754, #714821, #719571, #723168, #725541

[-0.4.9.42]
- Add 0089-RHBZ-694602-RSSM-config.patch
- Add 0090-RHBZ-700169-fix-nr-active.patch
   * Make sure to always count ghost paths as active
- Add 0091-RHBZ-699577-manpage-clarification.patch
- Add 0092-RHBZ-689504-rdac-retry.patch
   * Make the rdac checker retry up to 5 times on errors that are likely
     transient.
- Add 0093-RHBZ-677449-dont-remove-map-on-enomem.patch
   * Don't remove a multipath device if dm_map_present fails due to lack of
     memory
- Add 0094-RHBZ-707560-check-return-value.patch
   * Don't try to update path groups on a path priority change if the
     multipath device has disappeared.
- Add 0095-RHBZ-678673-no-path-groups.patch
   * Allow multipath devices with no path groups
- Add 0096-RHBZ-683616-ioship-support.patch
   * Modify the rdac checker to work better storage in IOSHIP mode
- Add 0097-RHBZ-697386-fix-shutdown-crash.patch
   * Add missing test for thread cancellation to the checker thread.
- Add 0098-RHBZ-706555-dont-update-pgs-in-manual.patch
   * When failback is set to manual, do not automatically reconfigure the
     pgs when a path changes priority
- Add 0099-RHBZ-705854-warn-on-bad-dev-loss-tmo.patch
- Add 0100-RHBZ-710478-deprecate-uid-gid-mode.patch
   * uid, gid and mode are now controlled by udev.
- Add 0101-RHBZ-631009-disable-udev-disk-rules-on-reload.patch
- Remove 0082-RHBZ-631009-disable-udev-disk-rules-on-reload.patch
   * multipath and kpartx now have a "-u" option that forces the udev
     dm-disk rules to run.
- Resolves: bz #694602, #700169, #689504, #677449, #707560, #678673
- Resolves: bz #683616, #697386, #706555, #705854, #710478, #631009

[-0.4.9.41]
- Add 0088-RHBZ-693524-fix-prio-segfault.patch
    * Don't reuse the the path argument as a variable in the vector
      loop.
- Resolves: bz# 696157

[-0.4.9.40]
- Add 0087-RHBZ-680480-skip-if-no-sysdev.patch
    * Don't check the path if it has no sysdev, since it is about to
      be removed.
- Resoves: bz# 680480

[-0.4.9.39]
- Add 0086-RHBZ-681144-sysfs-device-cleanup.patch
    * Make sure to remove the sysfs device from cache when the path is
      removed. Also, only search the sysfs device cache under the vecs lock
- Resolves: bz #681144

[-0.4.9.38]
- Remove 0082-RHBZ-631009-disable-udev-disk-rules-on-reload.patch
    * This patch was causing a serious regression (677937), and resolving
      it requires changes in /etc/rc.sysinit, so I'm backing it out for
      now.
- Resolves: bz #677937
- Related: bz #631009

[-0.4.9.37]
- Add 0085-RHBZ-645605-fix-offline-check.patch
    * Make path_offline return false if it can't deterime a paths state.
- Rename 0080-RHBZ-622731-fix-no-config-value-segfault.patch to
          0080-RHBZ-662731-fix-no-config-value-segfault.patch
    * accidentally used the wrong bug number
- Resolves: bz #645605, #662731

[-0.4.9.36]
- Add 0084-RHBZ-644111-read-only-bindings.patch
     * Add -B multipath and multipathd option, to make the bindings file
       read-only. This is to be used by the initramfs
- Resolves: bz #644111

[-0.4.9.35]
- Add 0082-RHBZ-631009-disable-udev-disk-rules-on-reload.patch
     * Set DM_UDEV_DISABLE_DISK_RULES_FLAG in dm_simplecmd to stop blkid 
from
       running on reloads
- Add 0083-RHBZ-636213-633643-new-configs.patch
     * Autoconfigs for HP P2000 family and NEC Storage M Series arrays
- Resolves: bz #631009, #636213, #636213

[-0.4.9.34]
- Add 0081-RHBZ-623644-fix-sysfs-caching.patch
     * Remove attribute value caching, and free cached parent devices on 
remove
- Resolves: bz #623644

[-0.4.9.33]
- Add 0073-RHBZ-650664-clarify-error-msg.patch
     * tell users to increase max_fds, if multipath gets EMFILE trying to
       open a file.
- Add 0074-RHBZ-602883-dont-print-change.patch
     * don't print messages for every change uevent
- Add 0075-RHBZ-576919-log-checker-err.patch
     * add log_checker_err default config option, to control the verbosity
       of repeated checker error messages
- Add 0076-RHBZ-599690-update-multipath-conf.patch
- Add 0077-RHBZ-622608-nvdisk-config.patch
- Add 0078-RHBZ-628095-config-warnings.patch
     * print warning messages for invalid config parameters
- Add 0079-RHBZ-650797-display-iscsi-tgt-name.patch
     * get tgt name for iscsi devices
- Add 0080-RHBZ-622731-fix-no-config-value-segfault.patch
     * check that a config option has a value before trying to get the
       strlen of the value.
- Resolves: bz #576919, #599690, #602883, #622608, #622731, #628095, #650664
- Resolves: bz #650797

[-0.4.9.32]
- Add 0064-RHBZ-612173-fix-reverse-lookup.patch
     * don't terminate WWID at whitespace on reverse lookup.
- Add 0065-RHBZ-635088-update-priority.patch
     * refresh all priorities when a new path comes online
- Add 0066-RHBZ-636071-mpathconf-variable_names.patch
     * don't use DISPLAY variable name. It configs with X. Also, unset all
       the variable names, so this doesn't happen again.
- Add 0067-RHBZ-622569-symmetrix-config.patch
- Add 0068-RHBZ-632734-nvdisk-config.patch
- Add 0069-RHBZ-636246-hp-open-config.patch
- Add 0070-RHBZ-639037-hitachi-open-config.patch
- Add 0071-RHBZ-611779-fix-whitespace-crash.patch
     * get_cmdvec wasn't setting the passed in vector to NULL, when
       there was no command.
- Add 0072-RHBZ-651389-change-scsi-tmo-order.patch
     * Make multipath set fast_io_fail_tmo before dev_loss_tmo, so that
       you can set dev_loss_tmo above 600.
- Resolves: bz #651389, #611779, #639037, #636246, #632734, #622569, #636071
- Resolves: bz #635088, #612173

[-0.4.9.31]
- Modify 0063-RHBZ-595719-udev_link_priority.patch
     * make link_priority only work on multipath devices and partitions.
- Resolves: bz #595719

[-0.4.9.30]
- Modify 0063-RHBZ-595719-udev_link_priority.patch
     * move link_priority line up in the udev rules. Hopefully this is the
       actual fix.
- Resolves: bz #595719

[-0.4.9.29]
- Modify 0063-RHBZ-595719-udev_link_priority.patch
     * Change link priority from 5 to 10
- Resolves: bz #595719

[-0.4.9.28]
- Add 0064-RHBZ-612173-fix-reverse-lookup.patch
    * Fix multipath's wwid bindings file reverse lookup
- Resolves: bz #612173

[-0.4.9.27]
- Add 0063-RHBZ-595719-udev_link_priority.patch
    * Increase the priority of multipath symlinks
- Resolves: bz #595719

[-0.4.9.26]
- Add 0062-RHBZ-592998-hpsc-config.patch
- Resolves: bz #592998

[-0.4.9.25]
- Add 0061-RHBZ-620479-find-rport.patch
   * multipath now will correctly set the dev_loss_tmo and fast_io_fail_tmo
     for setups where a devices rport id doesn't match its target id.
- Resolves: bz #620479

[-0.4.9.24]
- Modify 0055-RHBZ-602257-update-on-show-topology.patch
   * fix patch so that multipath only syncs with the kernel, and doesn't
     update other features
- Add 0060-RHBZ-606420-fix-remove-map.patch
   * multipathd will now only report a success for remove map when it
     actually succeeds.
- Resolves: bz #602257, #606420






More information about the El-errata mailing list