REGRESSSION in pending 4.7.4

classic Classic list List threaded Threaded
9 messages Options
Reply | Threaded
Open this post in threaded view
|

REGRESSSION in pending 4.7.4

Samba - samba-technical mailing list
Just a more clear warning that there is a regression in startup
handling in v4-7-test for which I have a patch under build in master.

See the  [PATCH] Run 'samba' daemon in foreground thread.  Please
cherry-pick that patch into 4.7.4 or revert the current patch that
landed before releasing 4.7.4.

Additionally, 4.7.3 introduced a regression in DNS for anything but the
 smallest of DNS zones.  I have patches pending in master for a partial
fix and will finish the final fix tomorrow (just need to adjust tests
to lock in a new, correct behaviour in LDB).  It would be nice to get
that in too, but I understand that waiting for a fix isn't desirable
either.

Thanks,

Andrew Bartlett
--
Andrew Bartlett                       http://samba.org/~abartlet/
Authentication Developer, Samba Team  http://samba.org
Samba Developer, Catalyst IT          http://catalyst.net.nz/services/samba


Reply | Threaded
Open this post in threaded view
|

DELAYING 4.7.4 (was: REGRESSSION in pending 4.7.4)

Samba - samba-technical mailing list
Hi,

On 19.12.2017 08:51, Andrew Bartlett wrote:

> Just a more clear warning that there is a regression in startup
> handling in v4-7-test for which I have a patch under build in master.
>
> See the  [PATCH] Run 'samba' daemon in foreground thread.  Please
> cherry-pick that patch into 4.7.4 or revert the current patch that
> landed before releasing 4.7.4.
>
> Additionally, 4.7.3 introduced a regression in DNS for anything but the
>  smallest of DNS zones.  I have patches pending in master for a partial
> fix and will finish the final fix tomorrow (just need to adjust tests
> to lock in a new, correct behaviour in LDB).  It would be nice to get
> that in too, but I understand that waiting for a fix isn't desirable
> either.

ok, thanks for the hint. Please update the bug report accordingly (the
4.7 patch should be marked obsolete, right?).

Delaying 4.7.4 until patches will be good to go.

Is there a bug report for the DNS issue?

Thanks!

Cheers,
Karo

--
Karolin Seeger https://samba.org/~kseeger/
Release Manager Samba Team https://samba.org
Team Lead Samba SerNet https://sernet.de

Reply | Threaded
Open this post in threaded view
|

Re: DELAYING 4.7.4 (was: REGRESSSION in pending 4.7.4)

Samba - samba-technical mailing list
On Tue, 2017-12-19 at 12:04 +0100, Karolin Seeger wrote:

> Hi,
>
> On 19.12.2017 08:51, Andrew Bartlett wrote:
> > Just a more clear warning that there is a regression in startup
> > handling in v4-7-test for which I have a patch under build in master.
> >
> > See the  [PATCH] Run 'samba' daemon in foreground thread.  Please
> > cherry-pick that patch into 4.7.4 or revert the current patch that
> > landed before releasing 4.7.4.
> >
> > Additionally, 4.7.3 introduced a regression in DNS for anything but the
> >  smallest of DNS zones.  I have patches pending in master for a partial
> > fix and will finish the final fix tomorrow (just need to adjust tests
> > to lock in a new, correct behaviour in LDB).  It would be nice to get
> > that in too, but I understand that waiting for a fix isn't desirable
> > either.
>
> ok, thanks for the hint. Please update the bug report accordingly (the
> 4.7 patch should be marked obsolete, right?).

It just needs one more patch from master as I won the autobuild lottery
last night and the patch landed. :-)

> Delaying 4.7.4 until patches will be good to go.
>
> Is there a bug report for the DNS issue?

https://bugzilla.samba.org/show_bug.cgi?id=13191

I'll tidy this all up today.

Thanks for your patience.

Andrew Bartlett
--
Andrew Bartlett                       http://samba.org/~abartlet/
Authentication Developer, Samba Team  http://samba.org
Samba Developer, Catalyst IT          http://catalyst.net.nz/services/samba


Reply | Threaded
Open this post in threaded view
|

Re: DELAYING 4.7.4 (was: REGRESSSION in pending 4.7.4)

Samba - samba-technical mailing list
On Wed, 2017-12-20 at 07:22 +1300, Andrew Bartlett via samba-technical
wrote:

> On Tue, 2017-12-19 at 12:04 +0100, Karolin Seeger wrote:
> > Hi,
> >
> > On 19.12.2017 08:51, Andrew Bartlett wrote:
> > > Just a more clear warning that there is a regression in startup
> > > handling in v4-7-test for which I have a patch under build in master.
> > >
> > > See the  [PATCH] Run 'samba' daemon in foreground thread.  Please
> > > cherry-pick that patch into 4.7.4 or revert the current patch that
> > > landed before releasing 4.7.4.
> > >
> > > Additionally, 4.7.3 introduced a regression in DNS for anything but the
> > >  smallest of DNS zones.  I have patches pending in master for a partial
> > > fix and will finish the final fix tomorrow (just need to adjust tests
> > > to lock in a new, correct behaviour in LDB).  It would be nice to get
> > > that in too, but I understand that waiting for a fix isn't desirable
> > > either.
> >
> > ok, thanks for the hint. Please update the bug report accordingly (the
> > 4.7 patch should be marked obsolete, right?).
>
> It just needs one more patch from master as I won the autobuild lottery
> last night and the patch landed. :-)
>
> > Delaying 4.7.4 until patches will be good to go.
> >
> > Is there a bug report for the DNS issue?
>
> https://bugzilla.samba.org/show_bug.cgi?id=13191
>
> I'll tidy this all up today.
>
> Thanks for your patience.

From my perspective this is all done and I'm just waiting for patches
to be applied to 4.7.  However metze raised an objection so I'll need
to wait to confirm he is happy with my explanation.

Thanks,

Andrew Bartlett

--
Andrew Bartlett                       http://samba.org/~abartlet/
Authentication Developer, Samba Team  http://samba.org
Samba Developer, Catalyst IT          http://catalyst.net.nz/services/samba


Reply | Threaded
Open this post in threaded view
|

Re: DELAYING 4.7.4 (was: REGRESSSION in pending 4.7.4)

Samba - samba-technical mailing list
In reply to this post by Samba - samba-technical mailing list
On Wed, 2017-12-20 at 07:22 +1300, Andrew Bartlett via samba-technical
wrote:

> On Tue, 2017-12-19 at 12:04 +0100, Karolin Seeger wrote:
> > Hi,
> >
> > On 19.12.2017 08:51, Andrew Bartlett wrote:
> > > Just a more clear warning that there is a regression in startup
> > > handling in v4-7-test for which I have a patch under build in master.
> > >
> > > See the  [PATCH] Run 'samba' daemon in foreground thread.  Please
> > > cherry-pick that patch into 4.7.4 or revert the current patch that
> > > landed before releasing 4.7.4.
> > >
> > > Additionally, 4.7.3 introduced a regression in DNS for anything but the
> > >  smallest of DNS zones.  I have patches pending in master for a partial
> > > fix and will finish the final fix tomorrow (just need to adjust tests
> > > to lock in a new, correct behaviour in LDB).  It would be nice to get
> > > that in too, but I understand that waiting for a fix isn't desirable
> > > either.
> >
> > ok, thanks for the hint. Please update the bug report accordingly (the
> > 4.7 patch should be marked obsolete, right?).
>
> It just needs one more patch from master as I won the autobuild lottery
> last night and the patch landed. :-)
>
> > Delaying 4.7.4 until patches will be good to go.
> >
> > Is there a bug report for the DNS issue?
>
> https://bugzilla.samba.org/show_bug.cgi?id=13191
>
> I'll tidy this all up today.
>
> Thanks for your patience.

Can I please plead for this now to be released?

I respect Metze's right to NACK the patch as a fellow team member, but
I must disagree in the strongest possible terms.  

We have a set of working, tested patches that address a critical issue
on production networks, but sadly no 4.7 release that can be deployed. 

Sadly:

- 4.7.0 contains the issue in REPLACE of group memberships, a security
bug and a file corruption issue.
- 4.7.1 additionally introduced the DNS wildcard feature and so made
all DNS queries impossibly slow on large zones
- 4.7.2 fixes the file corruption issue, but still has the other above
deficiencies
- 4.7.3 fixes the security issue, still leaving the above issues

Therefore we need 4.7.4 to be released to fix the issue in REPLACE of
group memberships, and to address the performance issue introduced with
the DNS wildcard code.

Remember, even with no wildcard records DNS performance is unusable.

We have numerous, real-world complaints regarding this on the samba
mailing list.

Finally, as you know we are trying to get users to run Samba 4.7
because of the missing read locks in Samba 4.6 and below.  Having no
acceptable 4.7 release is a major issue for a number of our larger
production networks.

Therefore I must implore that the developed, validated patches in BUG
13191 be un-NACKed and 4.7.4 released.

Thanks,

Andrew Bartlett
--
Andrew Bartlett                       http://samba.org/~abartlet/
Authentication Developer, Samba Team  http://samba.org
Samba Developer, Catalyst IT          http://catalyst.net.nz/services/samba


Reply | Threaded
Open this post in threaded view
|

Re: DELAYING 4.7.4

Samba - samba-technical mailing list
Hi,

On 21.12.2017 19:29, Andrew Bartlett wrote:

> On Wed, 2017-12-20 at 07:22 +1300, Andrew Bartlett via samba-technical
> wrote:
>> On Tue, 2017-12-19 at 12:04 +0100, Karolin Seeger wrote:
>>> Hi,
>>>
>>> On 19.12.2017 08:51, Andrew Bartlett wrote:
>>>> Just a more clear warning that there is a regression in startup
>>>> handling in v4-7-test for which I have a patch under build in master.
>>>>
>>>> See the  [PATCH] Run 'samba' daemon in foreground thread.  Please
>>>> cherry-pick that patch into 4.7.4 or revert the current patch that
>>>> landed before releasing 4.7.4.
>>>>
>>>> Additionally, 4.7.3 introduced a regression in DNS for anything but the
>>>>  smallest of DNS zones.  I have patches pending in master for a partial
>>>> fix and will finish the final fix tomorrow (just need to adjust tests
>>>> to lock in a new, correct behaviour in LDB).  It would be nice to get
>>>> that in too, but I understand that waiting for a fix isn't desirable
>>>> either.
>>>
>>> ok, thanks for the hint. Please update the bug report accordingly (the
>>> 4.7 patch should be marked obsolete, right?).
>>
>> It just needs one more patch from master as I won the autobuild lottery
>> last night and the patch landed. :-)
>>
>>> Delaying 4.7.4 until patches will be good to go.
>>>
>>> Is there a bug report for the DNS issue?
>>
>> https://bugzilla.samba.org/show_bug.cgi?id=13191
>>
>> I'll tidy this all up today.
>>
>> Thanks for your patience.
>
> Can I please plead for this now to be released?
>
> I respect Metze's right to NACK the patch as a fellow team member, but
> I must disagree in the strongest possible terms.  
>
> We have a set of working, tested patches that address a critical issue
> on production networks, but sadly no 4.7 release that can be deployed. 
>
> Sadly:
>
> - 4.7.0 contains the issue in REPLACE of group memberships, a security
> bug and a file corruption issue.
> - 4.7.1 additionally introduced the DNS wildcard feature and so made
> all DNS queries impossibly slow on large zones
> - 4.7.2 fixes the file corruption issue, but still has the other above
> deficiencies
> - 4.7.3 fixes the security issue, still leaving the above issues
>
> Therefore we need 4.7.4 to be released to fix the issue in REPLACE of
> group memberships, and to address the performance issue introduced with
> the DNS wildcard code.
>
> Remember, even with no wildcard records DNS performance is unusable.
>
> We have numerous, real-world complaints regarding this on the samba
> mailing list.
>
> Finally, as you know we are trying to get users to run Samba 4.7
> because of the missing read locks in Samba 4.6 and below.  Having no
> acceptable 4.7 release is a major issue for a number of our larger
> production networks.
>
> Therefore I must implore that the developed, validated patches in BUG
> 13191 be un-NACKed and 4.7.4 released.

Metze, please comment.

Thanks!

Cheers,
Karo

--
Karolin Seeger https://samba.org/~kseeger/
Release Manager Samba Team https://samba.org
Team Lead Samba SerNet https://sernet.de

Reply | Threaded
Open this post in threaded view
|

Re: DELAYING 4.7.4

Samba - samba-technical mailing list
Am 21.12.2017 um 21:07 schrieb Karolin Seeger:

> Hi,
>
> On 21.12.2017 19:29, Andrew Bartlett wrote:
>> On Wed, 2017-12-20 at 07:22 +1300, Andrew Bartlett via samba-technical
>> wrote:
>>> On Tue, 2017-12-19 at 12:04 +0100, Karolin Seeger wrote:
>>>> Hi,
>>>>
>>>> On 19.12.2017 08:51, Andrew Bartlett wrote:
>>>>> Just a more clear warning that there is a regression in startup
>>>>> handling in v4-7-test for which I have a patch under build in master.
>>>>>
>>>>> See the  [PATCH] Run 'samba' daemon in foreground thread.  Please
>>>>> cherry-pick that patch into 4.7.4 or revert the current patch that
>>>>> landed before releasing 4.7.4.
>>>>>
>>>>> Additionally, 4.7.3 introduced a regression in DNS for anything but the
>>>>>  smallest of DNS zones.  I have patches pending in master for a partial
>>>>> fix and will finish the final fix tomorrow (just need to adjust tests
>>>>> to lock in a new, correct behaviour in LDB).  It would be nice to get
>>>>> that in too, but I understand that waiting for a fix isn't desirable
>>>>> either.
>>>>
>>>> ok, thanks for the hint. Please update the bug report accordingly (the
>>>> 4.7 patch should be marked obsolete, right?).
>>>
>>> It just needs one more patch from master as I won the autobuild lottery
>>> last night and the patch landed. :-)
>>>
>>>> Delaying 4.7.4 until patches will be good to go.
>>>>
>>>> Is there a bug report for the DNS issue?
>>>
>>> https://bugzilla.samba.org/show_bug.cgi?id=13191
>>>
>>> I'll tidy this all up today.
>>>
>>> Thanks for your patience.
>>
>> Can I please plead for this now to be released?
>>
>> I respect Metze's right to NACK the patch as a fellow team member, but
>> I must disagree in the strongest possible terms.  
>>
>> We have a set of working, tested patches that address a critical issue
>> on production networks, but sadly no 4.7 release that can be deployed. 
>>
>> Sadly:
>>
>> - 4.7.0 contains the issue in REPLACE of group memberships, a security
>> bug and a file corruption issue.
>> - 4.7.1 additionally introduced the DNS wildcard feature and so made
>> all DNS queries impossibly slow on large zones
>> - 4.7.2 fixes the file corruption issue, but still has the other above
>> deficiencies
>> - 4.7.3 fixes the security issue, still leaving the above issues
>>
>> Therefore we need 4.7.4 to be released to fix the issue in REPLACE of
>> group memberships, and to address the performance issue introduced with
>> the DNS wildcard code.
>>
>> Remember, even with no wildcard records DNS performance is unusable.
>>
>> We have numerous, real-world complaints regarding this on the samba
>> mailing list.
>>
>> Finally, as you know we are trying to get users to run Samba 4.7
>> because of the missing read locks in Samba 4.6 and below.  Having no
>> acceptable 4.7 release is a major issue for a number of our larger
>> production networks.
>>
>> Therefore I must implore that the developed, validated patches in BUG
>> 13191 be un-NACKed and 4.7.4 released.
>
> Metze, please comment.
I removed the NACK, I still don't like it, but I guess it the best
option we currently have.

Should we release 4.7.4 with or without fixes for
the critical bugs?:
https://bugzilla.samba.org/show_bug.cgi?id=13197
"SMB2 close/lock/logoff can generate NT_STATUS_NETWORK_SESSION_EXPIRED"
and
https://bugzilla.samba.org/show_bug.cgi?id=13195
"g_lock conflict detection broken when processing stale entries"

Both can lead to data corruption or a service outage.

metze


signature.asc (853 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: DELAYING 4.7.4

Samba - samba-technical mailing list
On Thu, Dec 21, 2017 at 10:13:36PM +0100, Stefan Metzmacher via samba-technical wrote:

>
> I removed the NACK, I still don't like it, but I guess it the best
> option we currently have.
>
> Should we release 4.7.4 with or without fixes for
> the critical bugs?:
> https://bugzilla.samba.org/show_bug.cgi?id=13197
> "SMB2 close/lock/logoff can generate NT_STATUS_NETWORK_SESSION_EXPIRED"
> and
> https://bugzilla.samba.org/show_bug.cgi?id=13195
> "g_lock conflict detection broken when processing stale entries"
>
> Both can lead to data corruption or a service outage.

With those fixes, assuming my autobuild goes through today.

I'll review the back ports asap.

Jeremy.

Reply | Threaded
Open this post in threaded view
|

Re: DELAYING 4.7.4

Samba - samba-technical mailing list
In reply to this post by Samba - samba-technical mailing list
On Thu, 2017-12-21 at 22:13 +0100, Stefan Metzmacher wrote:

> I removed the NACK, I still don't like it, but I guess it the best
> option we currently have.

Thank you.  I do appreciate it.

Andrew Bartlett
--
Andrew Bartlett
https://samba.org/~abartlet/
Authentication Developer, Samba Team         https://samba.org
Samba Development and Support, Catalyst IT  
https://catalyst.net.nz/services/samba