anyone dealt with this? * ClusterContainsContainerInstancesException: The Cluster cannot be deleted while Container Instances are active.
would depends_on affect order of destruction?
phinze
ryangg: yep it should. the dependencies chain is walked backwards for destruction
ryangg
sweet I'll try that. maybe killing the nodes first will make it easier to kill the ecs cluster. :)
phinze
yep - if you run into any hiccups feel free to open an issue with some config snippets - always happy to talk something out on an issue thread
bichonfrise74 has quit
jacl joined the channel
ryangg
phinze: that worked brilliantly and now I can destroy ECS clusters. Thanks!
phinze
hooray!
phinze heads out for the night o/
ajw0100 joined the channel
potatoes_ joined the channel
btobolaski has quit
rmenn joined the channel
rmenn has quit
plucas joined the channel
em-dash has quit
em-dash joined the channel
KenDawg2 has quit
carlosdp joined the channel
em-dash has quit
rafik has quit
svij_ has quit
namra joined the channel
ryangg has quit
KenDawg2 joined the channel
KenDawg2 has quit
rgl joined the channel
kapluni joined the channel
kingryan joined the channel
kingryan
has anyone else seen a bug where you get a error about an aws security group having a dependent object, even though that security group should not be getting any updates?
kapluni
hey folks. trying out terraform for the first time, and loving it. has anyone had an issue trying to create an AWS elasticache cluster? i get `Error creating Elasticache: InvalidParameterCombination: Cluster can only be created in a VPC` - is this generally the case for terraform with elasticache?
rgl has quit
knite joined the channel
jacl has quit
KenDawg2 joined the channel
KenDawg2 has quit
never mind - just using the wrong instance type
KenDawg2 joined the channel
kingryan has quit
ajw0100 has quit
knite has quit
knite joined the channel
semarj joined the channel
potatoes_ joined the channel
kapluni has quit
potatoes_ has quit
kapluni_ joined the channel
kapluni_ has quit
razorfront joined the channel
carlosdp has quit
carlosdp joined the channel
carlosdp joined the channel
carlosdp has quit
carlosdp joined the channel
blahdodo joined the channel
jacl joined the channel
knite
hm...I'm having some trouble with a couple of route table associations. I'm doing a typical public/private subnet split with the correspondng routes. the public routes come up fine, but the private routes fail.
Various other resource creation bits removed, these should be the only relevant bits.
tphummel joined the channel
tphummel has quit
semarj has quit
a13x212 joined the channel
mbrevoort joined the channel
_jesse_
knite: could you pastebin an example config?
nothing jumps out from the output
tphummel joined the channel
maybe look at the route table associations that were set up and see if anything seems off
rmenn joined the channel
knite
_jesse_: comparing against someone else's terraform config post, I tentatively think that one route table for two public subnets is okay, but I need a separate table per private subnet?
at any rate, that's the main difference between my config and this example config. I don't understand why that would be necessary, but I'm trying it...
tphummel has quit
_jesse_
I think one route table should be fine
knite
I write my config in YAML - are you cool with a YAML paste, or would you prefer the JSON I pass to terraform?
tphummel joined the channel
_jesse_
JSON would probably be preferrable just so I can try to run it
jacl has quit
knite
hokay, that's a bit trickier, what with separate variable file, etc. let me gin up a minimal repro I can share.
_jesse_
that would be perfect, I can eyeball your YAML, but it'd be easiest to see what's up with something I can execute
knite
_jesse_: okay, got a trimmed down repro. this uses us-east-1, I can pull up the AMI for a different region if you need it. all you'll need is an existing key pair name you can use, and being cool with spinning up a t2.micro for a minute or 5. and even that can probably be replaced with an internet gateway and still repro.