Closed Bug 807341 Opened 12 years ago Closed 12 years ago

Test B2G flashing with BMM

Categories

(Infrastructure & Operations :: RelOps: General, task)

task
Not set
normal

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: dustin, Assigned: dividehex)

References

Details

From clint: ---- One thing I need you (Dustin) and Ted to do, is to test the BMM flashing of the panda B2G build. The kernel that has been installed on chassis 3 in bug 797629 should be flashable. But we need to test that. Ideally I'd like to test these flash mechanisms with these pandas (in order from most important to least): * B2G -> B2G I have uploaded an image (the system.img, userdata.img etc) to http://people.mozilla.org/~ctalbert/b2g-earlyprintk/ so you can test this. * Android -> B2G: this is next important on the list so we can understand how easy (or difficult) it is to flash something with our existig android linaro image to B2G. You can use the same b2g build for flashing. * B2G -> Android: while this is not super important (I don't anticipate we'll do this often, if ever), it is probably necessary so you can test Android->B2G.
Assignee: server-ops-releng → dustin
The pandas in chassis 3 are imaged differently from panda-0021, which is what I used to make the DHCP configuration in bug 800048, on the assumption that it had the "final" uboot image on it. Come to find out, there *is* no uboot image, so we'll need one of those. And then we'll probably need to manually re-image the sdcards in chassis 3 with that uboot image.
Depends on: 807428
Blocks: 802317
Jake, does our work today constitute a successful test?
Assignee: dustin → jwatkins
(In reply to Dustin J. Mitchell [:dustin] from comment #2) > Jake, does our work today constitute a successful test? YES. I would say your tests were successful. Once the sdcard was imaged with the uboot preseed image, I was able to deploy pre-seed->Android, pre-seed->B2G, Android->B2G, B2G->Android, B2G->B2G, and Android->android. Keep in mind the B2G process only installs a static build at this time. We will need to enhance the b2g second stage with API hooks to retrieve a deployment config so it may install different builds from buildbot.
WOOT! Y'all rock! Nice. What else needs to be done here to call this one done? It sounds to me like: * We need the uboot image somewhere the pandas can get to it. * We need mozpool to pull the builds out of buildbot and feed them into this system so that we can flash with arbitrary B2G builds(perhaps those are your API hooks you mentioned?)
Neither is really part of this bug. The uboot image is put onto the sdcards by dcops and its working bits are never overwritten. It's also already available: http://puppetagain.pub.build.mozilla.org/data/bmm/preseed/ As for mozpool and Buildbot interacting, you haven't correctly characterized that interaction, but yes, we need to finish up that high-level part of mozpool (request handling) and its interface to buildbot. Mark is working on that. We've been sprinting so quickly on mozpool that lots of the development hasn't been in bugs, but maybe that deserves one. Mark, Clint, open one if you need to? Anyway, we're done here.
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
Component: Server Operations: RelEng → RelOps
Product: mozilla.org → Infrastructure & Operations
You need to log in before you can comment on or make changes to this bug.