Chris Armstrong carmstrong

Organizations

@engineyard @deis
carmstrong commented on issue deis/deis#4281
@carmstrong

there was no v1.9.1 to checkout in trunk Really? That's certainly odd - we've tagged it just like any other release.

carmstrong commented on pull request deis/deis#4373
@carmstrong

Code LGTM.

carmstrong commented on pull request deis/deis#4374
@carmstrong

Code LGTM.

carmstrong commented on pull request deis/deis#4315
@carmstrong

Reminder: we need to switch path.cwd for something else (maybe path.root if it exists?) so that we can use the script root to find user-data. /cc @…

carmstrong commented on pull request deis/deis#4370
@carmstrong

Code LGTM.

@carmstrong
@carmstrong
  • @carmstrong 06afd1c
    docs(contrib/azure): update to use ARM
carmstrong commented on pull request deis/deis#4366
@carmstrong

Code LGTM.

carmstrong opened pull request deis/deis#4365
@carmstrong
[WIP] ref(contrib/azure): use ARM template for provisioning
2 commits with 416 additions and 334 deletions
carmstrong commented on issue deis/deis#4334
@carmstrong

I tried working around this by dropping the inboundNatRules and instead assigned public IP addresses to the hosts. I haven't been able to test this…

@carmstrong

Could you also please sign the CLA? Thanks! I signed mine. Not sure if my company also has to sign one?

carmstrong commented on pull request deis/deis#4363
@carmstrong

Code LGTM.

@carmstrong
carmstrong commented on pull request coreos/coreos-vagrant#254
@carmstrong

I just noticed that we're specifying "off" while the .box uses "on"...

carmstrong commented on issue deis/deis#4360
@carmstrong

CoreOS removed it It's actually still there, but baked into the CoreOS .box image rather than the Vagrantfile - see the discussion on coreos/core…

carmstrong commented on pull request deis/deis#4341
@carmstrong

:ship:

carmstrong commented on issue deis/deis#4358
@carmstrong

@daanemanz If you follow the fleet daemon (journalctl -fu fleetd) on the host that controller is scheduled to, what you do see when you try to star…

carmstrong commented on issue deis/deis#4354
@carmstrong

digital ocean Ah, okay. Other providers give us the ability to separate the etcd journal and Docker volume and place them on different devices. W…

carmstrong commented on issue deis/deis#4354
@carmstrong

@bf4 What cloud provider are you on? Did you provision using our provision scripts? And can you show me a df -h and a mount on one of your hosts?

carmstrong commented on pull request deis/rigger#1
@carmstrong

LGTM.

carmstrong commented on issue deis/deis#4355
@carmstrong

this should be safe to do every time the controller starts, right? I think so, but @bacongobbler or @mboersma are probably best to confirm.

carmstrong commented on issue deis/deis#4351
@carmstrong

(I'm looking froward to uninstalling tons of stuff from my Mac!) :beer:

carmstrong commented on pull request kubernetes/kubernetes#6649
@carmstrong

Yes, rebase please. We'd love to get this merged and shipped into k8s ASAP.

carmstrong commented on issue deis/deis#4354
@carmstrong

@bf4 Also, you're running a version of CoreOS (717.3.0) that is newer than our tested and recommended version of 647.2.0. There are known issues wi…

carmstrong commented on issue deis/deis#4354
@carmstrong

@bf4 Your errors look like etcd is falling over... what does journalctl -fu etcd look like on the deis-v173-1 host? Sorry to hear you're considerin…

carmstrong commented on pull request deis/deis.io#124
@carmstrong

We could also just remove the ASCIIcast since it's linked to from the GitHub issue. If users are interested, they can click through and watch?

carmstrong commented on issue deis/deis#4355
@carmstrong

Good idea. We already keep track of what needs to get done here: http://docs.deis.io/en/latest/managing_deis/backing_up_data/#finishing-up It shoul…

carmstrong commented on pull request deis/deis#4315
@carmstrong

Ah - looks like the path for user-data I'm using is using the cwd, not the directory the script is in. I bet if you first changed into contrib/digi…

@carmstrong

We could, although I figured that if we're asking the user to log in to grab their token, they may as well get their fingerprint.