Lindsay Holmwood <[email protected]> writes:
> On 19 May 2010 15:39, Luke Kanies <[email protected]> wrote:
>> On May 18, 2010, at 6:08 PM, Trevor Vaughan wrote:
>>
>>> -----BEGIN PGP SIGNED MESSAGE-----
>>> Hash: SHA1
[...]
>>> This got me thinking about the concept of atomic puppet updates. It
>>> shouldn't be too difficult to write to disk/register the state of the
>>> operations as they happen and to be able to pick back up by default if a
>>> run is interrupted.
[...]
>> This wouldn't be all that difficult - just record each event as it comes in,
>> and remove it once it's dealt with. Then when Puppet starts up, it just
>> deals with any un-dealt-with event, as it were.
>>
>> We currently only queue events in memory, but it's not like it's a huge
>> architectural shift to record the events on disk.
>
> This is very similar to how AngryMob[0] works internally, albiet without the
> serialisation to disk.
>
> I'm guessing the constant serialisation/de-serialisation would induce a
> relatively minor slowdown, but I think the feature is useful enough to
> warrant it.
If you expect this to be robust, it is going to require database style
checkpointing to disk to give solid assurances that the state is recorded.
Otherwise you lose if "interrupted" is, say, "a reboot".
Daniel
Sometimes, painfully, that is the root cause of this sort of interruption on
our network. You can't fix every daft sysadmin in one shot. :(
--
✣ Daniel Pittman ✉ [email protected] ☎ +61 401 155 707
♽ made with 100 percent post-consumer electrons
--
You received this message because you are subscribed to the Google Groups
"Puppet Developers" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/puppet-dev?hl=en.
