Re: [poky] RFC: design of network based PR service

Re: [poky] RFC: design of network based PR service

[Koen Kooi]

Adding oe-core list to CC:

Op 28 apr 2011, om 09:08 heeft Lu, Lianhao het volgende geschreven:

> Hi Guys,
> 
> Here is the design of network based PR service, please help to review and 
> give you comment. Thanks a lot!
> 
> By changing the BB_SIGNATURE_HANDLER from "basic" to "basichash", the 
> task checksum will be in the stamp file, so this implies the end of PR bumps 
> since whenever a recipe change happens, the checksum of the appropriate 
> tasks will change so everything depends on that tasks will automatically got 
> rebuilt. (The checksum mechanism was introduced here at https://lists.yoctoproject.org/pipermail/yocto/2011-March/001157.html).
> 
> But there is one piece of building block missing there. Because the package 
> feeds needs the PR bumps because some relevant fields in the package file(i.e. 
> Version field in .ipk and .deb file, Release field in .rpm file) need it so the 
> generated package feeds can be used to upgrade/install new packages in the 
> user's linux distribution. So RP proposed this network based PR service at 
> https://lists.yoctoproject.org/pipermail/poky/2011-January/002123.html. 
> 
> Each embedded linux distribution entity using Yocto to build their distributions 
> should have a single network based PR service. Multiple build sources could send 
> the tuple (PN, task checksum) to this network based PR service. If the PR service 
> finds the corresponding (PN, task checksum) in its database, it will return the PR 
> values to the client. If not, it will bump the PR for that PN and return the new value 
> to the client. The returned value will only goes into the generated package files.
> 
> The PR network service will be a separate process other than bitbake, using the 
> XMLPRC to communicate with the client(i.e. bitbake process) and sqlite3 as the 
> backend database. The table definition would be something like
> (
>  PN TEXT,
>  CheckSum TEXT PRIMARY KEY,
>  PR INT
> )
> 
> Because different type of package files are generated in different tasks, (e.g.
> do_package_write_ipk, do_package_write_rpm), we will use the task checksum 
> of do_package to ask for PR value, so all the package files will have the same PR value.
> 
> To be backward compatible, if the PR value is already set in the recipe, the final PR value 
> that goes into the package files will be "PR in recipe.PR returned by the service".
> 
> ========================
> Lianhao Lu
> OTC SSG, Intel APAC R&D Ltd.
> 86-21-61167139
> 
> 
> 
> 
> 
> _______________________________________________
> poky mailing list
> poky@yoctoproject.org
> https://lists.yoctoproject.org/listinfo/poky

Re: [poky] RFC: design of network based PR service

[Martin Jansa]

[-- Attachment #1: Type: text/plain, Size: 3440 bytes --]

On Thu, Apr 28, 2011 at 03:08:03PM +0800, Lu, Lianhao wrote:
> Hi Guys,
> 
> Here is the design of network based PR service, please help to review and 
> give you comment. Thanks a lot!

Hi,

looks good, just wondering if we can use the same mechanism for
LOCALCOUNT numbers in SRCPV, we can even use same table if we prefix
first column ie LOCALCOUNT_${PN} and CheckSum in 2nd column would be
actuall git hash (from SRCREV or HEAD for AUTOREV) or 2nd table with
similar structure.

But it wont work very well if one builder is using AUTOREV and another
one isn't, but that can be resolved by allowing only trusted builders
with same configuration (wrt AUTOREV) to send such tuples.

Do we have at least 2 groups of "users".
1) trusted which increments PR when hash is not found
2) public which can query PR for given hash (not sure what they should
   do if hash is not found)

Regards,

> By changing the BB_SIGNATURE_HANDLER from "basic" to "basichash", the 
> task checksum will be in the stamp file, so this implies the end of PR bumps 
> since whenever a recipe change happens, the checksum of the appropriate 
> tasks will change so everything depends on that tasks will automatically got 
> rebuilt. (The checksum mechanism was introduced here at https://lists.yoctoproject.org/pipermail/yocto/2011-March/001157.html).
> 
> But there is one piece of building block missing there. Because the package 
> feeds needs the PR bumps because some relevant fields in the package file(i.e. 
> Version field in .ipk and .deb file, Release field in .rpm file) need it so the 
> generated package feeds can be used to upgrade/install new packages in the 
> user's linux distribution. So RP proposed this network based PR service at 
> https://lists.yoctoproject.org/pipermail/poky/2011-January/002123.html. 
> 
> Each embedded linux distribution entity using Yocto to build their distributions 
> should have a single network based PR service. Multiple build sources could send 
> the tuple (PN, task checksum) to this network based PR service. If the PR service 
> finds the corresponding (PN, task checksum) in its database, it will return the PR 
> values to the client. If not, it will bump the PR for that PN and return the new value 
> to the client. The returned value will only goes into the generated package files.
> 
> The PR network service will be a separate process other than bitbake, using the 
> XMLPRC to communicate with the client(i.e. bitbake process) and sqlite3 as the 
> backend database. The table definition would be something like
> (
>   PN TEXT,
>   CheckSum TEXT PRIMARY KEY,
>   PR INT
> )
> 
> Because different type of package files are generated in different tasks, (e.g.
> do_package_write_ipk, do_package_write_rpm), we will use the task checksum 
> of do_package to ask for PR value, so all the package files will have the same PR value.
> 
> To be backward compatible, if the PR value is already set in the recipe, the final PR value 
> that goes into the package files will be "PR in recipe.PR returned by the service".
> 
> ========================
> Lianhao Lu
> OTC SSG, Intel APAC R&D Ltd.
> 86-21-61167139
> 
> 
> 
> 
> 
> _______________________________________________
> poky mailing list
> poky@yoctoproject.org
> https://lists.yoctoproject.org/listinfo/poky

-- 
Martin 'JaMa' Jansa     jabber: Martin.Jansa@gmail.com

Re: [poky] RFC: design of network based PR service

[Mark Hatle]

(adding oe-core back to the list as it got dropped off my previous reply)

On 4/28/11 4:04 PM, Joshua Lock wrote:
> On Thu, 2011-04-28 at 10:19 -0500, Mark Hatle wrote:
>> On 4/28/11 4:22 AM, Martin Jansa wrote:
>>> On Thu, Apr 28, 2011 at 03:08:03PM +0800, Lu, Lianhao wrote:
>>>> Hi Guys,
>>>>
>>>> Here is the design of network based PR service, please help to review and 
>>>> give you comment. Thanks a lot!
>>>
>>> Hi,
>>>
>>> looks good, just wondering if we can use the same mechanism for
>>> LOCALCOUNT numbers in SRCPV, we can even use same table if we prefix
>>> first column ie LOCALCOUNT_${PN} and CheckSum in 2nd column would be
>>> actuall git hash (from SRCREV or HEAD for AUTOREV) or 2nd table with
>>> similar structure.
>>>
>>> But it wont work very well if one builder is using AUTOREV and another
>>> one isn't, but that can be resolved by allowing only trusted builders
>>> with same configuration (wrt AUTOREV) to send such tuples.
>>>
>>> Do we have at least 2 groups of "users".
>>> 1) trusted which increments PR when hash is not found
>>> 2) public which can query PR for given hash (not sure what they should
>>>    do if hash is not found)
>>
>> I think this points our something necessary.  There are manual indications of a
>> change.  I.e. the current "PR" usage.  If I change the recipe, patches, etc I
>> should expect to update the 'PR' to the next value.  However, if something ELSE
>> changes (affecting the checksum), or an end user forgets to change the PR, the
>> auto increment should be used.
> 
> Why would you need to update the PR manually? Detecting changes in the
> recipe/patches/etc should all be handled by the checksumming.

Checksums and timestamps are not enough to determine a logical progression of
the components.

We need something that informs the user where these changes fit in the grand
scheme of things.  Are they newer or older then a recipe of the same name (and
package version)?

So this really allows us to visualize multiple levels of upstream work:

Level 1 - Upstream "version"  (PV)

Level 2 - Upstream (oe-core) recipe versioning (PR)

...

Level N - Local build and configuration versioning (checksum based & auto
incrementing)


In a lot of projects 1, 2 and N are enough.. but it's understandable to add in
other intermediate levels to indicate different upstreams along the way.  (Such
as if a company has their own internal distribution..)

I've seen cases where a level 3 exists to indicate a "distribution version"..

Level N (when done manually) has usually captured major system configuration
changes and rebuilding in order to enable solver tools to upgrade properly.


While the checksum will be able to point a unique instance of the recipe to a
given PR... it doesn't guaranty any type of logical numbering.. other then a new
checksum is a new (presumably larger) PR number.  By adding in the various
levels of version information the checksum becomes "more" unique as it only
refers to specific 'upstream' revisions, and make it more logical that a level
1, 2, ... N change means it's a "newer" version of the package.

--Mark

> Cheers,
> Joshua

Re: [poky] RFC: design of network based PR service

[Chris Larson]

On Thu, Apr 28, 2011 at 2:23 PM, Mark Hatle <mark.hatle@windriver.com> wrote:
> Checksums and timestamps are not enough to determine a logical progression of
> the components.
>
> We need something that informs the user where these changes fit in the grand
> scheme of things.  Are they newer or older then a recipe of the same name (and
> package version)?

I think we might want to stop using the term PR to describe what
you're talking about here.  PR has historically had a quite specific
meaning to us, given how bitbake has operated, and how stamps worked.
It sounds like you want to formalize what we've likely all been doing
manually -- PR .= ".1" or whatever in the .bbappend of a given layer.
Do you think we really need a format string for this, or would
introducing a new variable that's simply a list of extra version
components, and which is used by the packaging classes, likely not by
bitbake itself, get the job done? Am I grasping your need correctly?
-- 
Christopher Larson
clarson at kergoth dot com
Founder - BitBake, OpenEmbedded, OpenZaurus
Maintainer - Tslib
Senior Software Engineer, Mentor Graphics

Re: [poky] RFC: design of network based PR service

[Lu, Lianhao]

Mark Hatle wrote on 2011-04-29:
> (adding oe-core back to the list as it got dropped off my previous reply)
> 
> On 4/28/11 4:04 PM, Joshua Lock wrote:
>> On Thu, 2011-04-28 at 10:19 -0500, Mark Hatle wrote:
>>> On 4/28/11 4:22 AM, Martin Jansa wrote:
>>>> On Thu, Apr 28, 2011 at 03:08:03PM +0800, Lu, Lianhao wrote:
>>>>> Hi Guys,
>>>>> 
>>>>> Here is the design of network based PR service, please help to
>>>>> review and give you comment. Thanks a lot!
>>>> 
>>>> Hi,
>>>> 
>>>> looks good, just wondering if we can use the same mechanism for
>>>> LOCALCOUNT numbers in SRCPV, we can even use same table if we
>>>> prefix first column ie LOCALCOUNT_${PN} and CheckSum in 2nd column
>>>> would be actuall git hash (from SRCREV or HEAD for AUTOREV) or 2nd
>>>> table with similar structure.
>>>> 
>>>> But it wont work very well if one builder is using AUTOREV and
>>>> another one isn't, but that can be resolved by allowing only
>>>> trusted builders with same configuration (wrt AUTOREV) to send such tuples.
>>>> 
>>>> Do we have at least 2 groups of "users".
>>>> 1) trusted which increments PR when hash is not found
>>>> 2) public which can query PR for given hash (not sure what they should
>>>>    do if hash is not found)
>>> 
>>> I think this points our something necessary.  There are manual
>>> indications of a change.  I.e. the current "PR" usage.  If I change
>>> the recipe, patches, etc I should expect to update the 'PR' to the
>>> next value.  However, if something ELSE changes (affecting the
>>> checksum), or an end user forgets to change the PR, the auto
>>> increment
> should be used.
>> 
>> Why would you need to update the PR manually? Detecting changes in
>> the recipe/patches/etc should all be handled by the checksumming.
> 
> Checksums and timestamps are not enough to determine a logical
> progression of the components.
> 
> We need something that informs the user where these changes fit in the
> grand scheme of things.  Are they newer or older then a recipe of the
> same name (and package version)?
> 
> So this really allows us to visualize multiple levels of upstream work:
> 
> Level 1 - Upstream "version"  (PV)
> 
> Level 2 - Upstream (oe-core) recipe versioning (PR)
> 
> ...
> 
> Level N - Local build and configuration versioning (checksum based &
> auto incrementing)
> 
> 
> In a lot of projects 1, 2 and N are enough.. but it's understandable
> to add in other intermediate levels to indicate different upstreams along the way.
> (Such as if a company has their own internal distribution..)
> 
> I've seen cases where a level 3 exists to indicate a "distribution version"..
> 
> Level N (when done manually) has usually captured major system
> configuration changes and rebuilding in order to enable solver tools to upgrade properly.
> 
> 
> While the checksum will be able to point a unique instance of the
> recipe to a given PR... it doesn't guaranty any type of logical
> numbering.. other then a new checksum is a new (presumably larger) PR
> number.  By adding in the various levels of version information the
> checksum becomes "more" unique as it only refers to specific
> 'upstream' revisions, and make it more logical that a level 1, 2, ... N change means it's a "newer" version of the package.
> 

If a new PV(and/or PR) value comes, does the level N value need to be reset to 0? Or should it continue to increase?

Best Regards,
Lianhao

Re: [poky] RFC: design of network based PR service

[Khem Raj]

On Thu, Apr 28, 2011 at 10:28 PM, Lu, Lianhao <lianhao.lu@intel.com> wrote:
> Mark Hatle wrote on 2011-04-29:
>> (adding oe-core back to the list as it got dropped off my previous reply)
>>
>> On 4/28/11 4:04 PM, Joshua Lock wrote:
>>> On Thu, 2011-04-28 at 10:19 -0500, Mark Hatle wrote:
>>>> On 4/28/11 4:22 AM, Martin Jansa wrote:
>>>>> On Thu, Apr 28, 2011 at 03:08:03PM +0800, Lu, Lianhao wrote:
>>>>>> Hi Guys,
>>>>>>
>>>>>> Here is the design of network based PR service, please help to
>>>>>> review and give you comment. Thanks a lot!
>>>>>
>>>>> Hi,
>>>>>
>>>>> looks good, just wondering if we can use the same mechanism for
>>>>> LOCALCOUNT numbers in SRCPV, we can even use same table if we
>>>>> prefix first column ie LOCALCOUNT_${PN} and CheckSum in 2nd column
>>>>> would be actuall git hash (from SRCREV or HEAD for AUTOREV) or 2nd
>>>>> table with similar structure.
>>>>>
>>>>> But it wont work very well if one builder is using AUTOREV and
>>>>> another one isn't, but that can be resolved by allowing only
>>>>> trusted builders with same configuration (wrt AUTOREV) to send such tuples.
>>>>>
>>>>> Do we have at least 2 groups of "users".
>>>>> 1) trusted which increments PR when hash is not found
>>>>> 2) public which can query PR for given hash (not sure what they should
>>>>>    do if hash is not found)
>>>>
>>>> I think this points our something necessary.  There are manual
>>>> indications of a change.  I.e. the current "PR" usage.  If I change
>>>> the recipe, patches, etc I should expect to update the 'PR' to the
>>>> next value.  However, if something ELSE changes (affecting the
>>>> checksum), or an end user forgets to change the PR, the auto
>>>> increment
>> should be used.
>>>
>>> Why would you need to update the PR manually? Detecting changes in
>>> the recipe/patches/etc should all be handled by the checksumming.
>>
>> Checksums and timestamps are not enough to determine a logical
>> progression of the components.
>>
>> We need something that informs the user where these changes fit in the
>> grand scheme of things.  Are they newer or older then a recipe of the
>> same name (and package version)?
>>
>> So this really allows us to visualize multiple levels of upstream work:
>>
>> Level 1 - Upstream "version"  (PV)
>>
>> Level 2 - Upstream (oe-core) recipe versioning (PR)
>>
>> ...
>>
>> Level N - Local build and configuration versioning (checksum based &
>> auto incrementing)
>>
>>
>> In a lot of projects 1, 2 and N are enough.. but it's understandable
>> to add in other intermediate levels to indicate different upstreams along the way.
>> (Such as if a company has their own internal distribution..)
>>
>> I've seen cases where a level 3 exists to indicate a "distribution version"..
>>
>> Level N (when done manually) has usually captured major system
>> configuration changes and rebuilding in order to enable solver tools to upgrade properly.
>>
>>
>> While the checksum will be able to point a unique instance of the
>> recipe to a given PR... it doesn't guaranty any type of logical
>> numbering.. other then a new checksum is a new (presumably larger) PR
>> number.  By adding in the various levels of version information the
>> checksum becomes "more" unique as it only refers to specific
>> 'upstream' revisions, and make it more logical that a level 1, 2, ... N change means it's a "newer" version of the package.
>>
>
> If a new PV(and/or PR) value comes, does the level N value need to be reset to 0? Or should it continue to increase?
>

I am in favor of reseting it to 0

> Best Regards,
> Lianhao
>
>
> _______________________________________________
> poky mailing list
> poky@yoctoproject.org
> https://lists.yoctoproject.org/listinfo/poky
>