Merge ~paelzer/ubuntu/+source/multipath-tools:merge-artful into ~usd-import-team/ubuntu/+source/multipath-tools:debian/sid
| Status: | Merged |
|---|---|
| Merge reported by: | Robie Basak |
| Merged at revision: | cd2f5a906ca7f75c18f382f579b5e84ba07b68ea |
| Proposed branch: | ~paelzer/ubuntu/+source/multipath-tools:merge-artful |
| Merge into: | ~usd-import-team/ubuntu/+source/multipath-tools:debian/sid |
| Diff against target: |
1804 lines (+1434/-76) 19 files modified
debian/changelog (+1047/-0) debian/control (+15/-6) debian/initramfs/hooks (+22/-2) debian/initramfs/init-top (+18/-2) debian/initramfs/kpartx.hook (+20/-0) debian/initramfs/local-bottom (+62/-0) debian/initramfs/local-premount (+30/-0) debian/kpartx-boot.postinst (+34/-0) debian/kpartx-boot.postrm (+45/-0) debian/kpartx.install (+2/-2) debian/multipath-tools.dm-mpath-lvm.udev (+26/-0) debian/multipath-tools.preinst (+4/-0) debian/patches/1000--set-umask-in-multipathd.patch (+30/-0) debian/patches/disable-fexceptions-udeb.patch (+16/-0) debian/patches/enable-find-multipaths.patch (+11/-0) debian/patches/kpartx_more_loopback_fixes.patch (+40/-0) debian/patches/series (+4/-0) debian/rules (+8/-4) dev/null (+0/-60) |
| Related bugs: |
|
| Reviewer | Review Type | Date Requested | Status |
|---|---|---|---|
| Mathieu Trudel-Lapierre (community) | 2017-07-18 | Approve on 2017-07-18 | |
| LocutusOfBorg (community) | 2017-07-18 | Approve on 2017-07-18 | |
| Nish Aravamudan | Needs Fixing on 2017-07-18 | ||
|
Review via email:
|
|||
Description of the Change
Merge for Artful
| ChristianEhrhardt (paelzer) wrote : | # |
| Nish Aravamudan (nacc) wrote : | # |
W: Expected old/debian (f620d8a88361e0
E: Expected old/debian (41104278899b70
W: Expected old/ubuntu (571ebab5cf0020
E: Expected old/ubuntu (f4cd2bbb055d15
W: Expected new/debian (188a3d6b61d94c
E: Expected new/debian (83c617aa74241f
E: More than one changelog diff hunk detected
+multipath-tools (0.6.4-5ubuntu1) artful; urgency=medium
+
+ * Merge from Debian. (LP: #1548303) Remaining changes:
+ - d/control:
+ - multipath-udeb: add sg3-udeb Depends
+ - d/rules: Move udev rules to priority 95, because rules that load modules
+ should be >90.
+ - d/multipath-
+ that multipathd will be able to start.
+ - Split kpartx initramfs bits into kpartx-boot for dmraid (LP 941874)
+ - d/initramfs/
+ - d/kpartx-
+ - d/kpartx-
+ - d/control: Add kpartx-boot package for dmraid
+ - d/rules: Install kpartx initramfs hook
+ - d/kpartx.install: install all arch /lib* kpartx udev rules
+ - patches to multipath source
+ - d/p/1000-
+ - d/p/kpartx_
+ files some more: since we stat() the loopback device node, we can't rely
+ on S_ISREG() tests to handle this case, and should look at the device
+ itself instead. (LP 1543430)
+ - d/p/enable-
+ by default -- see the removed 'add_find-
+ - multipath initramfs fixes for booting from multipathed devices
+ - d/initramfs/hooks: also copy wwids file on the installed system to
+ ensure all paths come up on boot. (LP 1479929)
+ - d/initramfs/hooks: install multipathd and required directories.
+ - d/initramfs/hooks: copy dm-mpath-lvm & multipath udev rules to initramfs
+ - d/initramfs/hooks: do not copy kpartx rules to initramfs
+ - d/initramfs/
+ - d/initramfs/
+ resolve_device() in local_mount_root(), so the by-uuid/ symlinks have a
+ chance to be updated by the multipath udev rules (LP 1503286).
+ - d/initramfs/
+ nor change /etc/multipath/
+ - d/rules: install d/initramfs/
+ - d/rules: install d/initramfs...
| ChristianEhrhardt (paelzer) wrote : | # |
I only have a small testbed in regard to multipath, but did what I could to verify:
- s390x multipath setup (2 disks, 2 paths each, 1:1 mappings of adapter to wwn)
- upgrade to new package - ok
- reboot with new packaage - ok
- autopkgtests ran against the bileto ppa - ok
- this need sto run vs proposed as liburcu6 is currently in proposed migration
- FYI Build in Bileto: https:/
I think this is good, but lets wait a little bit for other opinions and for liburcu migration
| Mathieu Trudel-Lapierre (cyphermox) wrote : | # |
Looks fine to me, except maybe that now I wonder why we use partx in the udev rules rather than kpartx (I see no benefit to --nr... ); please don't think that I "own" multipath-tools though. You should feel free to upload and do the requisite testing, as much as you can, to validate that multipath-tools works correctly.
| ChristianEhrhardt (paelzer) wrote : | # |
Thanks for all your reviews.
I'm likely going to upload then, but I want to look into the reports by the automated linter, posting findings one by one and eventually conclude with fixes and/or please merge now.
E: More than one changelog diff hunk detected
That is odd, the report shows:
2017 14:12:53 +0200
multipath-tools (0.6.4-5) unstable; urgency=medium
* Ship multipath udev rules (Closes: #858456)
-- Patrick Caulfield <email address hidden> Wed, 3 Nov 2004 10:16:34 +0000
+
But the context there is way too old.
Maybe it got it wrong due to the old old/debian, new/debian, ... - more on that in the next post.
The following looks sane:
$ git diff new/debian -- debian/changelog
| ChristianEhrhardt (paelzer) wrote : | # |
W: Expected old/debian (f620d8a88361e0
E: Expected old/debian (41104278899b70
W: Expected old/ubuntu (571ebab5cf0020
E: Expected old/ubuntu (f4cd2bbb055d15
W: Expected new/debian (188a3d6b61d94c
E: Expected new/debian (83c617aa74241f
Yeah I didn't repush old/new branches to to my repo.
Not important for the actual merge as the merge base is correct.
But the linter found the old ones of the zesty merge still around.
I pushed the updated tags in my repo to match the current USDI state to fix it.
| ChristianEhrhardt (paelzer) wrote : | # |
There are two partx (instead of kpartx) references:
1. in debian/
copy_exec /usr/bin/partx
2. debian/
RUN+=
These reference are there "that way" since
+multipath-tools (0.5.0-7ubuntu15) xenial; urgency=medium
+
+ [ Mauricio Faria de Oliveira ]
+ * Remove partition device nodes of individual paths (for LVM on multipath)
+ (LP: #1540401)
+ - debian/
+ - debian/
+ - debian/
+ - debian/rules: install the udev rule (priority 56: after 55-scsi-sg3_id)
+ * debian/rules: don't ship 95-multipath.rules udev rules anymore; they are
+ not necessary with multipath-tools listening for udev events directly.
+
+ -- Mathieu Trudel-Lapierre <email address hidden> Thu, 11 Feb 2016 19:08:14 -0500
These files are part of multipath-tools which depends on kpartx.
So yeah one could change that, but that is not (doesn't have to be) part of the merge IMHO.
| ChristianEhrhardt (paelzer) wrote : | # |
I think we are ready to upload th emerge tag and I would upload the package.
But there is a new rebuild in artful-p that we need to integrate
https:/
I could just smash it into the changelog as it is CL only, but I might affect the merges to keep the rich history.
I'd more like:
1. an re-import of multipath
2. making sure changelogs are correct
3. I need to fixup and modify logical/
4. push all those tags
Hmm much effort for a pure changelog entry that we are going to overwrite in artful-proposed in a minute.
Can we ignore that if it was never release?
- cd2f5a9... by ChristianEhrhardt on 2017-07-19
| ChristianEhrhardt (paelzer) wrote : | # |
changelog: insert non-imported no change rebuilds
There are two no-change rebuilds in artful-proposed which were missing
on ubuntu/devel so far.
After a discussion on [1] we agreed that adding a new commit at the tail
of the merge is fine to keep rich history.
[1]: https:/
| ChristianEhrhardt (paelzer) wrote : | # |
Addressed all review feedback and ready to merge before upload.
| Nish Aravamudan (nacc) wrote : | # |
On Wed, Jul 19, 2017 at 10:10 AM, ChristianEhrhardt
<email address hidden> wrote:
> W: Expected old/debian (f620d8a88361e0
> E: Expected old/debian (41104278899b70
> W: Expected old/ubuntu (571ebab5cf0020
> E: Expected old/ubuntu (f4cd2bbb055d15
> W: Expected new/debian (188a3d6b61d94c
> E: Expected new/debian (83c617aa74241f
>
> Yeah I didn't repush old/new branches to to my repo.
> Not important for the actual merge as the merge base is correct.
> But the linter found the old ones of the zesty merge still around.
Right, the linter does not treat the above as fatal (F: ). But it does
indicate they are problems for the linter itself (it can't lint
further really without the above passing).
| Nish Aravamudan (nacc) wrote : | # |
On Wed, Jul 19, 2017 at 10:05 AM, ChristianEhrhardt
<email address hidden> wrote:
> Thanks for all your reviews.
> I'm likely going to upload then, but I want to look into the reports by the automated linter, posting findings one by one and eventually conclude with fixes and/or please merge now.
>
>
> E: More than one changelog diff hunk detected
>
> That is odd, the report shows:
> 2017 14:12:53 +0200
> multipath-tools (0.6.4-5) unstable; urgency=medium
>
> * Ship multipath udev rules (Closes: #858456)
>
> -- Patrick Caulfield <email address hidden> Wed, 3 Nov 2004 10:16:34 +0000
>
> +
>
>
> But the context there is way too old.
> Maybe it got it wrong due to the old old/debian, new/debian, ... - more on that in the next post.
I think this might be a bug in git-merge-
now. Basically, the linter takes new/debian and runs
`git-merge-
and compares that treeish to the proposed target to ensure only new
changelogs entries are introduced. For some reason, that treeish is
removing the last line in d/changelog relative to new/debian.


I set Mathieu as reviewer as he is the real owner of multipath and Locutus as he asked for the merge today.