Recommended update for mdadm

Announcement ID: SUSE-RU-2018:1331-1
Rating: moderate
References:
Affected Products:
  • SLES for SAP Applications 11-SP4
  • SUSE Linux Enterprise Server 11 SP4

An update that has 12 fixes can now be installed.

Description:

This update for mdadm provides the backporting of some critical fixes from upstream, and replace some existing patches with their upstream counterpart (bsc#1081910). The following fixes are included:

  • Grow: Followup fix for a problem that reshape fails to continue after restart on RAID array. (bsc#881530)
  • IMSM: Add warning message when assemble spanned container. (bsc#882634)
  • mdmon: Allow prepare_update to report failures. (bsc#1081910)
  • DDF, IMSM: Validate metadata_update size before using it. (bsc#1081910)
  • Grow: Do not try to restart if reshape is running. (bsc#887773)
  • IMSM: Move "validate_container_imsm" to be included in mdassemble. (bsc#1081910)
  • Grow: Fix a problem that was preventing the resize of an array to 32bit size. (bsc#1081910)
  • Assemble: Only fail auto-assemble in case of mdadm.conf conflicts. (bsc#1081910)
  • super: Make sure to ignore disk state flags that we don't understand. (bsc#1081910)
  • mdmon: Don't include super0 and super1 in mdmon. (bsc#1081910)
  • config: Add a new option to suppress adding bad block lists. (bsc#1081910)
  • Manage: Fix the removal of non-existent devices. (bsc#1081910)
  • Monitor: Stop monitoring devices that have disappeared. (bsc#1081910)
  • super1: Don't allow adding a bitmap if there is no space. (bsc#1081910)
  • super1: Make sure 'room' includes 'bbl_size' when creating array. (bsc#1081910)
  • update: Add 'bbl' and 'no-bbl' to the list of known updates. (bsc#1081910)
  • Grow: Report when grow needs metadata update. (bsc#1081910)
  • Grow: Fix the resize of array component size to > 32bits. (bsc#1081910)
  • mdcheck: Do not trigger an error if no /dev/md?* devices exists. (bsc#1081910)
  • Rebuildmap: Strip local host name from device name. (bsc#1081910)
  • Detail: Fix handling of 'disks' array. (bsc#1081910)
  • Incremental: Do not be distracted by partition table when calling try_spare. (bsc#1081910)
  • imsm: Add support for OROMs shared by multiple HBAs. (fate#317456)
  • imsm: Add support for second and combined AHCI controllers in UEFI mode. (fate#317456)
  • imsm: Add support for NVMe devices. (fate#317456)
  • imsm: Use efivarfs interface for reading UEFI variables. (fate#317456)
  • Monitor: Do not open md array that doesn't exist. (bsc#1081910)
  • mdcheck: Be careful when sourcing the output of "mdadm --detail --export". (bsc#1081910)
  • Monitor: Fix for regression with container devices. (bsc#1081910)
  • Grow.c: Fix classic readlink() buffer overflow. (bsc#1081910)
  • imsm: Simplified multiple OROMs support. (bsc#1081910)
  • IncRemove: Set "auto-read" only after successful excl open. (bsc#1081910)
  • Assemble: Fix "no uptodate device" message. (bsc#1081910)
  • Assemble: Revert the support for assembling of a RAID0 being reshaped. (bsc#1081910)
  • Assemble: Allow a RAID4 to assemble easily when parity devices is missing. (bsc#926767)
  • Assemble/force: Make it possible to "force" a new device in a reshape. (bsc#1081910)
  • IMSM: Count arrays per orom. (bsc#926517)
  • Add "Name" defines to some ancillary programs. (bsc#1081910)
  • Assemble: Don't check for pre-existing array when updating uuid. (bsc#1081910)
  • Manage: When re-adding, do check avail size if ->sb cannot be found. (bsc#1081910)
  • Grow: Only warn about incompatible metadata when no fallback available. (bsc#1081910)
  • Grow: Be more careful if array is stopped during critical section. (bsc#1081910)
  • Grow: Retry when writing 'reshape' to 'sync_action' is EBUSY. (bsc#1081910)
  • Grow: Be even more careful about handing a '0' completed value. (bsc#1081910)
  • Grow: Another attempt to fix stop-during-reshape race. (bsc#1081910)
  • Grow: Fix problem with --grow --continue. (bsc#1081910)
  • Create n bitmaps for clustered mode. (bsc#1081910)
  • Add nodes option while creating md. (bsc#1081910)
  • Set home-cluster while creating an array. (bsc#1081910)
  • Show all bitmaps while examining bitmap. (bsc#1081910)
  • Add a new clustered disk. (bsc#1081910)
  • Convert a bitmap=none device to clustered. (bsc#1081910)
  • Skip clustered devices in incremental. (bsc#1081910)
  • mdadm: Add the ability to change cluster name. (bsc#1081910)
  • Assemble: Ensure stripe_cache is big enough to handle new chunk size. (bsc#1081910)
  • mdstat: Discard 'dev' field, just use 'devnm'. (bsc#1081910)
  • sysfs: Reject reads that use the whole buffer. (bsc#1081910)
  • Monitor: Don't wait forever on a 'frozen' array. (bsc#1081910)
  • Manage/stop: Guard against 'completed' being too large. (bsc#1081910)
  • Manage/stop: Don't stop during initial critical section. (bsc#1081910)
  • raid6check: Report role of suspect device. (bsc#1081910)
  • raid6check: Get device ordering correct for syndrome calculation. (bsc#1081910)
  • restripe: Fix data block order in raid6_2_data_recov. (bsc#1081910)
  • Assemble: Extend --homehost='<ignore>' to allow --name= to ignore homehost. (bsc#1081910)
  • mdassemble: Add "Name" definition. (bsc#1081910)
  • mdassemble: Include mapfile support. (bsc#1081910)
  • super1: Do not create bad block log for clustered devices. (bsc#1081910)
  • Fix --incremental handling on cluster array. (bsc#1081910)
  • mdadm: Make cluster raid also support re-add. (bsc#1081910)
  • re-add: Make re-add try to write sysfs node first. (bsc#1081910)
  • imsm: Don't call abort_reshape() in imsm_manage_reshape(). (bsc#1081910)
  • Show device as journal in --detail --examine. (bsc#1081910)
  • Enable create array with write journal (--write-journal DEVICE). (bsc#1081910)
  • Assemble array with write journal. (bsc#1081910)
  • Check write journal in incremental. (bsc#1081910)
  • Safeguard against writing to an active device of another node. (bsc#1081910)
  • Make cmap_ also have same policy as dlm_. (bsc#1081910)
  • Add crc32c and use it for r5l checksum. (bsc#1081910)
  • Avoid confusion with parameter 'devname' with same name, ensure buffer is large enough for two ints plus extras. (bsc#1081910)
  • Make sure 'path' buffer is large enough to fit 200 characters plus null terminator. (bsc#1081910)
  • mdadm: Change timestamps to unsigned data type. (bsc#1081910)
  • Add sysfs_array_state to struct mdinfo. (bsc#1081910)
  • mdadm: Allow cluster raid to also add disk within incremental mode. (bsc#1081910)
  • mdadm: Don't show cluster name once the bitmap is cleared. (bsc#1081910)
  • mdadm: Do not display bitmap info if it is cleared. (bsc#1081910)
  • mdadm: Do not try to hold dlm lock in free_super1. (bsc#1081910)
  • mdadm: Improve the safeguard for change cluster raid's sb. (bsc#1081910)
  • Detail: Report correct raid-disk for removed drives. (bsc#1081910)
  • Move journal to end of --detail list. (bsc#1081910)
  • Check and remove bitmap first when reshape to raid0. (bsc#1081910)
  • Detail: Fix wrong condition in recent change. (bsc#1081910)
  • Grow: Close file descriptor earlier to avoid problems. (bsc#1081910)
  • Add casts for the addr argument of connect and bind. (bsc#1081910)
  • util: Fix wrong return value of cluster_get_dlmlock. (bsc#1081910)
  • super1: Fix calculation of space_before. (bsc#1081910)
  • systemd/mdadm-last-resort: Add Conflicts to .service file. (bsc#1081910)
  • super0: Fix reporting of devices between 2GB and 4GB. (bsc#1081910)
  • super1: Allow reshape that hasn't really started to be reverted. (bsc#1081910)
  • super1: Fix bblog_size accesses on big-ending machines. (bsc#953595)
  • Manage.c: Only issue change events for kernels older than 2.6.28. (bsc#1081910)
  • super-intel: Ensure suspended region is removed when reshape completes. (bsc#1081910)
  • Fix wrong bitmap output for cluster raid. (bsc#1081910)
  • load_sys(): Add a buffer size argument. (bsc#1081910)
  • Fix regression during add devices. (bsc#1081910)
  • Change the option from NoUpdate to NodeNumUpdate. (bsc#1081910)
  • mdadm:Add '--nodes' option in GROW mode. (bsc#1081910)
  • Grow: Handle failure to load superblock in Grow_addbitmap(). (bsc#1081910)
  • Remove: Container should wait for an array to release a drive. (bsc#1081910)
  • Fix RAID metadata check. (bsc#1081910)
  • super1: Make write_bitmap1 compatible with previous mdadm versions. (bsc#1081910)
  • Allow level migration only for single-array container. (bsc#1081910)
  • Fix bus error when accessing MBR partition records. (bsc#1081910)
  • Add function for getting member drive sector size. (bsc#1081910)
  • Add failfast support. (fate#311379)
  • Use disk sector size value to set offset for reading GPT. (bsc#1081910)
  • reshape: Support raid5 grow on certain older kernels. (bsc#923920)
  • super1: Fix sb->max_dev when adding a new disk in linear array. (bsc#1032802)
  • Fix a problem that was causing DASD not to be re-added automatically to the MD Array after channel path is online. (bsc#808647)
  • Ignore empty inactive arrays. (bsc#966773)
  • Update md_monitor from 5.2 to 5.4 to make sure md_monitor -cMirrorStatus detects md mirror status properly. (bsc#1068175)

Patch Instructions:

To install this SUSE update use the SUSE recommended installation methods like YaST online_update or "zypper patch".
Alternatively you can run the command listed for your product:

  • SUSE Linux Enterprise Server 11 SP4
    zypper in -t patch slessp4-mdadm-13610=1
  • SLES for SAP Applications 11-SP4
    zypper in -t patch slessp4-mdadm-13610=1

Package List:

  • SUSE Linux Enterprise Server 11 SP4 (s390x x86_64 i586 ppc64 ia64)
    • mdadm-3.3.1-10.18.1
  • SLES for SAP Applications 11-SP4 (ppc64 x86_64)
    • mdadm-3.3.1-10.18.1

References: