Re: v1 -> v2 migration tests

[Daniel Axtens]

Hi Stephen,

Thank you so much for such a detailed investigation! Very glad to hear
patchwork is doing the right things.

Regards,
Daniel

> On Thu, 2017-06-22 at 11:16 +0100, Stephen Finucane wrote:
>> On Thu, 2017-06-22 at 09:05 +1000, Daniel Axtens wrote:
>> > Hi Stephen,
>> > 
>> > > Thanks for doing this. I've tinkered with both instances and haven't seen
>> > > any
>> > > serious issues. I'm most happy with the fact that the performance of the
>> > > web UI
>> > > hasn't been altered (at least, from my brief testing).
>> > > 
>> > > I'm thinking we give this another day or two and we might be good to go?
>> > 
>> > Russell picked up one thing that looked questionable:
>> > https://py3.patchwork.dja.id.au/project/lkml/list/?series==
>> > =&;
>> > q=powerpc%2Fnuma%3A+update==
>> > 
>> > It looks like it's slightly misthreaded; I haven't had a chance to dig
>> > into it properly just yet. It could well be that the code is correct and
>> > the messages were just sent with weird headers.
>> 
>> I took a look at the v6 versions of these...
>
> I also just took a look at "v1" (though it doesn't have a prefix, of course)
> and v5 of this over lunch. The "v1" patches/covers weren't threaded, but 
> that's
> because they were missing references and in-reply-to headers. We've fixed this
> since those patches were merged, so were they to be parsed today that wouldn't
> happen. As for v5, from looking at marc.info [1], it would appear that patch
> 1/2 was never sent or never made it to the list. Patch 2/2 in this series is
> also missing 'References' and 'In-Reply-To' headers, but because the patches
> arrived *after* we'd fixed that issue, they're threaded correctly.
>
> Important metadata about the patches is provided below, but all in all, I 
> think
> Patchwork is working a-ok, which is good to hear.
>
> Stephen
>
> PS: I can't comment on the quality of the patches themselves, but this guy
> really needs to start sending his patches with git-send-email instead of
> Thunderbird :)
>
> [1] https://marc.info/?a=14570338841=1=2
>
> ---
>
> Original series ("v1"):
>
>   [0/2] powerpc/dlpar: Correct display of hot-add/hot-remove CPUs and memory
> Date: Tue, 23 May 2017 10:15:11 -0500
> Received: from vger.kernel.org (vger.kernel.org [209.132.180.67])
>   by patchwork.dja.id.au (Postfix) with ESMTP id 48630CA6D5
>   for ; Wed, 24 May 2017 01:15:45 +1000
> Message-Id: <790af26b-7055-7997-2080-f967aef2d...@linux.vnet.ibm.com>
>
> https://py3.patchwork.dja.id.au/cover/15561/
>
>   [1/2] powerpc/numa: Update CPU topology when VPHN enabled
> Date: Tue, 23 May 2017 10:15:29 -0500
> Received: from vger.kernel.org (vger.kernel.org [209.132.180.67])
>   by patchwork.dja.id.au (Postfix) with ESMTP id DE4FDCA6D5
>   for ; Wed, 24 May 2017 01:15:47 +1000
> Message-Id: <4a1bec9a-d3d2-c0bd-3956-e6e402be3...@linux.vnet.ibm.com>
>
> https://py3.patchwork.dja.id.au/patch/15562/
>
>   [2/2] : powerpc/hotplug/mm: Fix hot-add memory node assoc
> Date: Tue, 23 May 2017 10:15:44 -0500
> Received: from vger.kernel.org (vger.kernel.org [209.132.180.67])
>   by patchwork.dja.id.au (Postfix) with ESMTP id 222F2CA6D5
>   for ; Wed, 24 May 2017 01:16:03 +1000
> Message-Id: <3bb44d92-b2ff-e197-4bdf-ec6d588d6...@linux.vnet.ibm.com>
>
> https://py3.patchwork.dja.id.au/patch/15563/
>
> Fifth revision (v5):
>
>   [V5,0/2] powerpc/dlpar: Correct display of hot-add/hot-remove CPUs and 
> memory
> Date: Sun, 18 Jun 2017 13:45:17 -0500
> Received: from vger.kernel.org (vger.kernel.org [209.132.180.67])
>   by patchwork.dja.id.au (Postfix) with ESMTP id 2CDAFBE57D
>   for ; Mon, 19 Jun 2017 04:45:30 +1000
> Message-Id: <2a7f2a4f-5885-d186-2baf-da72ce071...@linux.vnet.ibm.com>
>
> https://py3.patchwork.dja.id.au/cover/27710/
>
> [V5,2/2] powerpc/numa: Update CPU topology when VPHN enabled
>   Date: Sun, 18 Jun 2017 13:46:55 -0500
>   Received: from vger.kernel.org (vger.kernel.org [209.132.180.67])
> by patchwork.dja.id.au (Postfix) with ESMTP id 13FEBCE81B
> for ; Mon, 19 Jun 2017 04:47:11 +1000
>   Message-Id: 
>
>   https://py3.patchwork.dja.id.au/patch/27711/
___
Patchwork mailing list
Patchwork@lists.ozlabs.org
https://lists.ozlabs.org/listinfo/patchwork

Re: v1 -> v2 migration tests

[Stephen Finucane]

On Thu, 2017-06-22 at 11:16 +0100, Stephen Finucane wrote:
> However, despite this, Patchwork seems to managing just fine. There's only
> one patch out of series, and that's the one that has been sent as 1/2
> mistakenly. I don't know why this didn't get lumped into the existing series
> but it's correct behavior so meh.

Found it:

  https://github.com/getpatchwork/patchwork/blob/v2.0.0-rc4/patchwork/parser.py
#L915

We actually codify this. Still seems correct.

Stephen

___
Patchwork mailing list
Patchwork@lists.ozlabs.org
https://lists.ozlabs.org/listinfo/patchwork

Re: v1 -> v2 migration tests

[Stephen Finucane]

On Thu, 2017-06-22 at 11:16 +0100, Stephen Finucane wrote:
> On Thu, 2017-06-22 at 09:05 +1000, Daniel Axtens wrote:
> > Hi Stephen,
> > 
> > > Thanks for doing this. I've tinkered with both instances and haven't seen
> > > any
> > > serious issues. I'm most happy with the fact that the performance of the
> > > web UI
> > > hasn't been altered (at least, from my brief testing).
> > > 
> > > I'm thinking we give this another day or two and we might be good to go?
> > 
> > Russell picked up one thing that looked questionable:
> > https://py3.patchwork.dja.id.au/project/lkml/list/?series==
> > =&;
> > q=powerpc%2Fnuma%3A+update==
> > 
> > It looks like it's slightly misthreaded; I haven't had a chance to dig
> > into it properly just yet. It could well be that the code is correct and
> > the messages were just sent with weird headers.
> 
> I took a look at the v6 versions of these...

I also just took a look at "v1" (though it doesn't have a prefix, of course)
and v5 of this over lunch. The "v1" patches/covers weren't threaded, but that's
because they were missing references and in-reply-to headers. We've fixed this
since those patches were merged, so were they to be parsed today that wouldn't
happen. As for v5, from looking at marc.info [1], it would appear that patch
1/2 was never sent or never made it to the list. Patch 2/2 in this series is
also missing 'References' and 'In-Reply-To' headers, but because the patches
arrived *after* we'd fixed that issue, they're threaded correctly.

Important metadata about the patches is provided below, but all in all, I think
Patchwork is working a-ok, which is good to hear.

Stephen

PS: I can't comment on the quality of the patches themselves, but this guy
really needs to start sending his patches with git-send-email instead of
Thunderbird :)

[1] https://marc.info/?a=14570338841=1=2

---

Original series ("v1"):

  [0/2] powerpc/dlpar: Correct display of hot-add/hot-remove CPUs and memory
Date: Tue, 23 May 2017 10:15:11 -0500
Received: from vger.kernel.org (vger.kernel.org [209.132.180.67])
  by patchwork.dja.id.au (Postfix) with ESMTP id 48630CA6D5
  for ; Wed, 24 May 2017 01:15:45 +1000
Message-Id: <790af26b-7055-7997-2080-f967aef2d...@linux.vnet.ibm.com>

https://py3.patchwork.dja.id.au/cover/15561/

  [1/2] powerpc/numa: Update CPU topology when VPHN enabled
Date: Tue, 23 May 2017 10:15:29 -0500
Received: from vger.kernel.org (vger.kernel.org [209.132.180.67])
  by patchwork.dja.id.au (Postfix) with ESMTP id DE4FDCA6D5
  for ; Wed, 24 May 2017 01:15:47 +1000
Message-Id: <4a1bec9a-d3d2-c0bd-3956-e6e402be3...@linux.vnet.ibm.com>

https://py3.patchwork.dja.id.au/patch/15562/

  [2/2] : powerpc/hotplug/mm: Fix hot-add memory node assoc
Date: Tue, 23 May 2017 10:15:44 -0500
Received: from vger.kernel.org (vger.kernel.org [209.132.180.67])
  by patchwork.dja.id.au (Postfix) with ESMTP id 222F2CA6D5
  for ; Wed, 24 May 2017 01:16:03 +1000
Message-Id: <3bb44d92-b2ff-e197-4bdf-ec6d588d6...@linux.vnet.ibm.com>

https://py3.patchwork.dja.id.au/patch/15563/

Fifth revision (v5):

  [V5,0/2] powerpc/dlpar: Correct display of hot-add/hot-remove CPUs and memory
Date: Sun, 18 Jun 2017 13:45:17 -0500
Received: from vger.kernel.org (vger.kernel.org [209.132.180.67])
  by patchwork.dja.id.au (Postfix) with ESMTP id 2CDAFBE57D
  for ; Mon, 19 Jun 2017 04:45:30 +1000
Message-Id: <2a7f2a4f-5885-d186-2baf-da72ce071...@linux.vnet.ibm.com>

https://py3.patchwork.dja.id.au/cover/27710/

[V5,2/2] powerpc/numa: Update CPU topology when VPHN enabled
  Date: Sun, 18 Jun 2017 13:46:55 -0500
  Received: from vger.kernel.org (vger.kernel.org [209.132.180.67])
by patchwork.dja.id.au (Postfix) with ESMTP id 13FEBCE81B
for ; Mon, 19 Jun 2017 04:47:11 +1000
  Message-Id: 

  https://py3.patchwork.dja.id.au/patch/27711/
___
Patchwork mailing list
Patchwork@lists.ozlabs.org
https://lists.ozlabs.org/listinfo/patchwork

Re: v1 -> v2 migration tests

[Stephen Finucane]

On Thu, 2017-06-22 at 09:05 +1000, Daniel Axtens wrote:
> Hi Stephen,
> 
> > Thanks for doing this. I've tinkered with both instances and haven't seen
> > any
> > serious issues. I'm most happy with the fact that the performance of the
> > web UI
> > hasn't been altered (at least, from my brief testing).
> > 
> > I'm thinking we give this another day or two and we might be good to go?
> 
> Russell picked up one thing that looked questionable:
> https://py3.patchwork.dja.id.au/project/lkml/list/?series===;
> q=powerpc%2Fnuma%3A+update==
> 
> It looks like it's slightly misthreaded; I haven't had a chance to dig
> into it properly just yet. It could well be that the code is correct and
> the messages were just sent with weird headers.

I took a look at the v6 versions of these - it looks to be some messed up
emailing and nothing more.

There are two issues:

- He's resent a series with the same version on different dates.
- He's sent patch 2/2 of the first series as 1/2 mistakenly, then sent it as
2/2 (with a broken diff, for that matter) immediately after.

However, despite this, Patchwork seems to managing just fine. There's only one
patch out of series, and that's the one that has been sent as 1/2 mistakenly. I
don't know why this didn't get lumped into the existing series but it's correct
behavior so meh.

I've given a break down of the patches below, along with important metadata, in
case anyone reads differently into it.

Stephen

---

First series (v6):

  [V6,0/2] powerpc/dlpar: Correct display of hot-add/hot-remove CPUs and memory
Date: Mon, 19 Jun 2017 17:09:20 -0500
Received: from vger.kernel.org (vger.kernel.org [209.132.180.67])
  by patchwork.dja.id.au (Postfix) with ESMTP id 25D50CE7E4
  for ; Tue, 20 Jun 2017 08:09:57 +1000
Message-Id: <081bb19d-3008-842d-6872-f51fccb94...@linux.vnet.ibm.com>

https://py3.patchwork.dja.id.au/cover/28644/

[V6,1/2] powerpc/hotplug: Ensure enough nodes avail for operations
  Date: Mon, 19 Jun 2017 17:10:20 -0500
  Received: from vger.kernel.org (vger.kernel.org [209.132.180.67])
by patchwork.dja.id.au (Postfix) with ESMTP id 0821BCE7E4
for ; Tue, 20 Jun 2017 08:11:11 +1000
  Message-Id: 
  In-Reply-To: <081bb19d-3008-842d-6872-f51fccb94...@linux.vnet.ibm.com>

  https://py3.patchwork.dja.id.au/patch/28647/


  !!! Start broken patch

[V6,1/2] powerpc/numa: Update CPU topology when VPHN enabled
  Date: Mon, 19 Jun 2017 17:10:29 -0500
  Received: from vger.kernel.org (vger.kernel.org [209.132.180.67])
by patchwork.dja.id.au (Postfix) with ESMTP id 10B4CCE820
for ; Tue, 20 Jun 2017 08:11:12 +1000
  Message-Id: <2b085918-be91-7654-9e1c-eacd9b8fa...@linux.vnet.ibm.com>
  In-Reply-To: <081bb19d-3008-842d-6872-f51fccb94...@linux.vnet.ibm.com>

  https://py3.patchwork.dja.id.au/patch/28648/

  !!! End broken patch

[V6,2/2] powerpc/numa: Update CPU topology when VPHN enabled
  Date: Mon, 19 Jun 2017 17:13:27 -0500
  Received: from vger.kernel.org (vger.kernel.org [209.132.180.67])
by patchwork.dja.id.au (Postfix) with ESMTP id A1EABBE57D
for ; Tue, 20 Jun 2017 08:13:41 +1000
  Message-Id: 
  In-Reply-To: <081bb19d-3008-842d-6872-f51fccb94...@linux.vnet.ibm.com>

  https://py3.patchwork.dja.id.au/patch/28649/

Second series (also v6, but a resend):

  [V6,0/2] powerpc/dlpar: Correct display of hot-add/hot-remove CPUs and memory
Date: Tue, 20 Jun 2017 10:14:23 -0500
Received: from vger.kernel.org (vger.kernel.org [209.132.180.67])
by patchwork.dja.id.au (Postfix) with ESMTP id B7FB9CEB1B
for ; Wed, 21 Jun 2017 01:15:39 +1000
Message-Id: b0e02344-965b-5298-4ef6-3ec67d120...@linux.vnet.ibm.com

https://py3.patchwork.dja.id.au/cover/29113/

[V6,1/2] powerpc/hotplug: Ensure enough nodes avail for operations
  Date: Tue, 20 Jun 2017 10:14:57 -0500
  Received: from vger.kernel.org (vger.kernel.org [209.132.180.67])
by patchwork.dja.id.au (Postfix) with ESMTP id E8EB2CEB1B
for ; Wed, 21 Jun 2017 01:15:27 +1000
  Message-Id: 
  In-Reply-To: 

  https://py3.patchwork.dja.id.au/patch/29112/

    [V6,2/2] powerpc/numa: Update CPU topology when VPHN enabled
  Date: Tue, 20 Jun 2017 10:15:00 -0500
  Received: from vger.kernel.org (vger.kernel.org [209.132.180.67])
by patchwork.dja.id.au (Postfix) with ESMTP id 657ACCEB56
for ; Wed, 21 Jun 2017 01:16:08 +1000
  Message-Id: 

Re: v1 -> v2 migration tests

[Daniel Axtens]

Hi Stephen,

> Thanks for doing this. I've tinkered with both instances and haven't seen any
> serious issues. I'm most happy with the fact that the performance of the web 
> UI
> hasn't been altered (at least, from my brief testing).
>
> I'm thinking we give this another day or two and we might be good to go?

Russell picked up one thing that looked questionable:
https://py3.patchwork.dja.id.au/project/lkml/list/?seriespowerpc%2Fnuma%3A+update==

It looks like it's slightly misthreaded; I haven't had a chance to dig
into it properly just yet. It could well be that the code is correct and
the messages were just sent with weird headers.

I'll probably get a moment to look at it in the next couple of days if
you don't get to it before then.

Regards,
Daniel

___
Patchwork mailing list
Patchwork@lists.ozlabs.org
https://lists.ozlabs.org/listinfo/patchwork

Re: v1 -> v2 migration tests

[Stephen Finucane]

On Wed, 2017-06-21 at 00:16 +1000, Daniel Axtens wrote:
> Hi all,
> 
> I am running tests of v1 to v2 migrations. The origin is
> v1.patchwork.dja.id.au and the destination is
> migrate.patchwork.dja.id.au.
> 
> Every hour, data is synced from 'v1' to 'migrate' - emails are only
> ingested in v1.
> 
> I haven't seen any _obvious_ issues, but please do check it out.
> 
> Please continue to test things, including your API clients. (Obviously
> events won't work!) Users created in v1 should exist in migrate (after
> the hourly sync).

Thanks for doing this. I've tinkered with both instances and haven't seen any
serious issues. I'm most happy with the fact that the performance of the web UI
hasn't been altered (at least, from my brief testing).

I'm thinking we give this another day or two and we might be good to go?

Stephen
___
Patchwork mailing list
Patchwork@lists.ozlabs.org
https://lists.ozlabs.org/listinfo/patchwork