Re: [PATCH 00/13] fix for GitHub issue #108 and misc cleanups

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Wed, 2025-01-22 at 22:16 -0500, Benjamin Marzinski wrote:
> This patchset has a couple of different pieces, but that main part
> is fixing the issue where, if a path is added while offline (for
> instance by doing a "multipath reconfigure" while the path is
> offline)
> It won't be able to get the SCSI device serial number (or its ALUA
> target port group). This meant that the GROUP_BY_SERIAL and
> GROUP_BY_TPG path grouping policies would not group the path
> correctly,
> even once it came back online.
> 
> Patches 1 & 2 are just a fix on top of Martin's recent device
> reloading
> work.
> 
> Patches 3, 4, & 5 are changes to need_reload. They avoid setting and
> checking it in cases where it won't change how multipathd behaves.
> The side benefit of doing this work is that it's easier to reason
> about
> what setting need_reload can and can't effect. Specifically, it will
> now
> only cause multipathd to fix incorrect device tables while in
> checker_finished().
> 
> Patches 6 - 11 deal with the serial and tpg_id problem from issue
> #108
> 
> Patch 12 is shutting up a static analyzer warning that popped up on a
> recent rhel build.
> 
> Patch 13 is a fix for an issue I ran into while testing this set.
> 
> Benjamin Marzinski (13):
>   multipathd: use symbolic values for wait_for_udev
>   multipathd: don't reload the map when waiting for udev
>   libmultipath: don't set need_reload in adopt_paths
>   libmultipath: remove need_reload check from select_action
>   libmultipath: drop return value from update_pathvec_from_dm
>   libmultipath: Dont scan pointless sysfs parents
>   libmultipath: move sysfs checks out of scsi_ioctl_pathinfo
>   libmultipath: rename DI_SERIAL to DI_IOCTL
>   libmultipath: retry DI_IOCTL functions if they were skipped
>   libmultipath: set need_reload when pp->serial gets set
>   libmultipath: set need_reload when the tpg_id gets set
>   libmultipath: stop static analyzer complaint in init_foreign
>   libmultipath: be lenient in allowing the alua-based pgpolicies
> 
>  libmultipath/configure.c   |   4 +-
>  libmultipath/devmapper.c   |  22 +++++---
>  libmultipath/discovery.c   | 104 ++++++++++++++++-------------------
> --
>  libmultipath/discovery.h   |   7 ++-
>  libmultipath/foreign.c     |   2 +-
>  libmultipath/propsel.c     |   6 ++-
>  libmultipath/structs.h     |  15 +++++-
>  libmultipath/structs_vec.c |  10 ++--
>  multipath/main.c           |   2 +-
>  multipathd/cli_handlers.c  |  16 +++---
>  multipathd/main.c          |  40 +++++++-------
>  tests/test-lib.c           |   9 +---
>  12 files changed, 119 insertions(+), 118 deletions(-)
> 

For the set:
Reviewed-by: Martin Wilck <mwilck@xxxxxxxx>






[Index of Archives]     [DM Crypt]     [Fedora Desktop]     [ATA RAID]     [Fedora Marketing]     [Fedora Packaging]     [Fedora SELinux]     [Yosemite Discussion]     [KDE Users]     [Fedora Docs]

  Powered by Linux