Add jroll for ironic ocata PTL
Change-Id: I8a5ef8dd497f9d459b6d884be92f0010ea04406b
This commit is contained in:
parent
4f58543c67
commit
ddccfa5096
78
candidates/ocata/Ironic/jroll.txt
Normal file
78
candidates/ocata/Ironic/jroll.txt
Normal file
@ -0,0 +1,78 @@
|
|||||||
|
Hi friends,
|
||||||
|
|
||||||
|
I'd like to be the ironic PTL for another cycle, if you'll have me.
|
||||||
|
|
||||||
|
Newton, in my opinion, has been Ironic's best cycle yet. We got a ton of work
|
||||||
|
done, and made good progress on the work we didn't finish. Some of the
|
||||||
|
highlights from this cycle include:
|
||||||
|
|
||||||
|
* Upgrade testing. We've struggled to make progress until this cycle, when we
|
||||||
|
really went all-out on getting it done. We also have multinode CI working
|
||||||
|
and awaiting reviews, which is a great thing to have anyway, and also a huge
|
||||||
|
first step to testing rolling upgrades.
|
||||||
|
|
||||||
|
* Multitenant networking. This has been our biggest feature request for years,
|
||||||
|
and we've been hacking on it since Liberty. We finally got it done (though
|
||||||
|
of course, there's more to add on top) in both Ironic and Nova this cycle.
|
||||||
|
|
||||||
|
* Speaking of Nova, we've made huge progress there. We now support multiple
|
||||||
|
nova-compute services running the Ironic driver. We have a nova-compatible
|
||||||
|
console done in Ironic, with patches ready for Nova to land at the beginning
|
||||||
|
of Ocata. We've laid the groundwork on both sides to make scheduling better,
|
||||||
|
and remove the (host, node) tuple that we plagued Nova with for years. And
|
||||||
|
I believe we've made large strides in how we interact with the Nova team;
|
||||||
|
things seem much more friendlier, and people are willing to help each other
|
||||||
|
out more on both sides.
|
||||||
|
|
||||||
|
* Fine-grained keystone policy. This is a huge step in helping operators
|
||||||
|
improve the security of their deployments.
|
||||||
|
|
||||||
|
* Our server and client both properly support Keystone v3 now.
|
||||||
|
|
||||||
|
* We now have third-party CI for iLO, iRMC, Cisco UCS/CIMC, OneView, PowerKVM,
|
||||||
|
and DRAC. We're on a path to removing drivers without CI. For context, a year
|
||||||
|
ago we had no third-party CI. :)
|
||||||
|
|
||||||
|
* ironic-python-agent is now the only supported deploy ramdisk, and the APIs
|
||||||
|
it uses have been promoted to first class APIs.
|
||||||
|
|
||||||
|
* Ironic now emits software metrics and has a framework for emitting
|
||||||
|
notifications for use by Searchlight and other systems.
|
||||||
|
|
||||||
|
* Existing systems may now be 'adopted' by ironic to help folks migrate their
|
||||||
|
existing environments to be managed by ironic.
|
||||||
|
|
||||||
|
There's many more things we've done in the 3 releases we shipped this cycle,
|
||||||
|
but this is a PTL nomination, not Newton release notes, so I'll stop there.
|
||||||
|
|
||||||
|
Going forward into Ocata, there's a few major things I'd like to focus on:
|
||||||
|
|
||||||
|
* Rolling upgrades. Most of the base IaaS has this or will have it soon. We're
|
||||||
|
on a good path to getting it done, and I think it's super important for
|
||||||
|
deployers.
|
||||||
|
|
||||||
|
* Continue progress on the networking work. We need to finish portgroups and
|
||||||
|
get a plan (and hopefully some code) together for vlan-aware baremetal
|
||||||
|
instances. This will allow an arbitrary number of Neutron networks to be
|
||||||
|
connected over an arbitrary number of physical NICs, which is incredibly
|
||||||
|
useful to users.
|
||||||
|
|
||||||
|
* Driver composition. This will get rid of our ever-increasing driver matrix
|
||||||
|
and help users make more sense of all the driver options. It also helps
|
||||||
|
clean up lots of internal code, so it's important to everyone.
|
||||||
|
|
||||||
|
* Boot from volume. This didn't make as much progress as I liked in Newton,
|
||||||
|
so we should keep working on it and try to at least get all the plumbing
|
||||||
|
landed, if not the first implementation. Lots of folks are asking for this,
|
||||||
|
and it's the best path we have to something that looks like 'nova migrate'.
|
||||||
|
|
||||||
|
* Above all else, I plan to make it a priority to mentor a few folks that could
|
||||||
|
be the next PTL. I didn't do a good job of this in Newton (though I think
|
||||||
|
some people could do the job now). I think it's important to rotate the
|
||||||
|
leadership, so I hope that the project is in a position where I can step
|
||||||
|
down after Ocata, and write more code in Pike.
|
||||||
|
|
||||||
|
Whether I'm elected PTL or not, I plan to be working upstream on ironic
|
||||||
|
full-time during Ocata, and look forward to another successful cycle. :)
|
||||||
|
|
||||||
|
// jim
|
Loading…
Reference in New Issue
Block a user