Commit Graph

237 Commits

Author SHA1 Message Date
Paul Hinze
6c93fbb85d core: [refactor] store Deposed resource instances as a list
Deposed instances need to be stored as a list for certain pathological
cases where destroys fail for some reason (e.g. upstream API failure,
Terraform interrupted mid-run). Terraform needs to be able to remember
all Deposed nodes so that it can clean them up properly in subsequent
runs.

Deposed instances will now never touch the Tainted list - they're fully
managed from within their own list.

Added a "multiDepose" test case that walks through a scenario to
exercise this.
2015-03-04 12:25:59 -06:00
Mitchell Hashimoto
11d073f7d4 terraform: test the increase from one case 2015-03-01 21:39:48 -08:00
Mitchell Hashimoto
2389251c38 terraform: catch scenario where both "foo" and "foo.0" are in state 2015-03-01 21:28:41 -08:00
Mitchell Hashimoto
ac6efa5e57 Merge pull request #1065 from hashicorp/f-taint
Add "taint" command
2015-02-26 23:55:20 -08:00
Paul Hinze
c03e44106f core: band-aid fix for tainted double destroy
After a lot of fun debugging with @mitchellh we finally have a diagnosis
for #1056.

I'm going to attempt to reproduce the diagnosis in prose to test out my
understanding.

----

The `DestroyTransformer` runs twice:

 * `DestroyPrimary` mode creates destroy nodes for normal resources
 * `DestroyTainted` mode creates destroy nodes for tainted resources

These destroy nodes are specializations of `GraphConfigNode`, which has
a `DynamicExpand` step.

In `DynamicExpand` we have a race condition between the normal destroy
node and the tainted destroy node for a given resource when
`CreateBeforeDestroy` is off.

The `DestroyTainted` `GraphConfigNode` must run the `TaintedTransform`
to draw out tainted nodes, since it is reponsible for destroying them
for replacement.

The `DestroyPrimary` `GraphConfigNode` _also_ runs `TaintedTransform` -
this is to catch `Deposed` nodes from CBD, which are piggy backing on
the end of the `Tainted` list.

Here's the bug: when CBD is off and you start an apply with a tainted
resource in your state, both `DestroyPrimary` and `DestroyTainted` catch
it and create their own `graphNodeTaintedResource` in their respective
subgraphs.

If parallelism is disabled, this doesn't happen because the
`DestroyPrimary` subgraph resolves completely before the
`DestroyTainted` node does its `DynamicExpand`, so the `Tainted` list
has been cleared by the time `DestroyTainted` is expanding. With
parallelism, each of these two subgraphs plays out in its own goroutine
simultaneously, and each picks up the tainted resource(s) that the apply
starts with.  So you get two `graphNodeTaintedResource` nodes, and two
destroys.

This band-aid fixes that by skipping the TaintedTransform alltogether in
the `DestroyPrimary` node if CBD is off.

A better fix will follow, which involves reworking the `Deposed` concept
so it no longer piggybacks on `Tainted`.

fixes #1056
2015-02-26 22:32:50 -06:00
Mitchell Hashimoto
341be226f4 terraform: test for various taint cases 2015-02-26 09:50:18 -08:00
Mitchell Hashimoto
ac167c3082 terraform: test post state update is called 2015-02-23 22:10:31 -08:00
Mitchell Hashimoto
dd7bc5db0c terraform: more self tests 2015-02-23 15:02:26 -08:00
Mitchell Hashimoto
0e59acc2c9 terraform: enable self vars 2015-02-23 14:56:02 -08:00
Mitchell Hashimoto
022967acdb terraform: module inputs/vars can be non-strings [GH-819] 2015-02-23 13:50:53 -08:00
Mitchell Hashimoto
40ee70d5c9 terraform: Input should only be called on providers once
/cc @sethvargo

Prior to this commit, we'd only persist the result of calling Input if
any input was given (len(result) > 0). The result was that every module
would also repeat asking for input even if there was no input to be
asked for.

This commit makes it so that if no input was received, we still set a
sentinel so that modules don't re-ask.
2015-02-20 15:35:57 -08:00
Mitchell Hashimoto
ab414de0ab terraform: test case for destroying with a count [GH-527] 2015-02-20 10:02:52 -08:00
Mitchell Hashimoto
4b6f258f1f terraform: only depose state if we're doing the destroy side 2015-02-19 12:08:34 -08:00
Mitchell Hashimoto
10e96ad72d terraform: provisioner config is properly interpolated 2015-02-19 12:08:33 -08:00
Mitchell Hashimoto
459ad04d71 terraform: provider cache should append the provider name to it 2015-02-19 12:08:33 -08:00
Mitchell Hashimoto
7ccba588ac terraform: gofmt -w -r 'Context2 -> Context' terraform/ 2015-02-19 12:08:32 -08:00
Mitchell Hashimoto
e37c187228 terraform: goodbye context.go (old) 2015-02-19 12:08:32 -08:00
Mitchell Hashimoto
14b5942453 terraform: all input tests pass 2015-02-19 12:08:08 -08:00
Mitchell Hashimoto
7c78a3749e terraform: provider input 2015-02-19 12:08:08 -08:00
Mitchell Hashimoto
f3714f1efc terraform: start Input 2015-02-19 12:08:08 -08:00
Mitchell Hashimoto
960ba73f1d terraform: last apply test 2015-02-19 12:08:08 -08:00
Mitchell Hashimoto
1c713878b0 terraform: fix outputs on destroy 2015-02-19 12:08:07 -08:00
Mitchell Hashimoto
bd0c3b12cb terraform: more tests passing 2015-02-19 12:08:07 -08:00
Mitchell Hashimoto
c8091aa338 terraform: fix more create-before-destroy state juggling 2015-02-19 12:08:07 -08:00
Mitchell Hashimoto
33cad6d207 terraform: fix create before destroy edge case with partial state 2015-02-19 12:08:07 -08:00
Mitchell Hashimoto
f7f1f17b49 terraform: create before destroy 2015-02-19 12:08:07 -08:00
Mitchell Hashimoto
119d5a09cf terraform: fix some diff comparison 2015-02-19 12:08:07 -08:00
Mitchell Hashimoto
5f8d1b86d5 terraform: more passing tests 2015-02-19 12:08:07 -08:00
Mitchell Hashimoto
6bb3cb8f2f terraform: a lot more passing tests 2015-02-19 12:08:06 -08:00
Mitchell Hashimoto
742b45886a terraform: destroy nodes work properly 2015-02-19 12:08:06 -08:00
Mitchell Hashimoto
c3df003624 terraform: more passing tests 2015-02-19 12:08:06 -08:00
Mitchell Hashimoto
e2abf17c9c terraform: provisioners should only be run on first create 2015-02-19 12:08:06 -08:00
Mitchell Hashimoto
c70cc682ea terraform: resource provisioner self references are not put in the graph 2015-02-19 12:08:06 -08:00
Mitchell Hashimoto
edfb11ff1d terraform: more passing tests 2015-02-19 12:08:06 -08:00
Mitchell Hashimoto
79746cb35d terraform: state must not be nil for apply 2015-02-19 12:08:06 -08:00
Mitchell Hashimoto
819aed67d4 terraform: provisioners 2015-02-19 12:08:06 -08:00
Mitchell Hashimoto
f8871917f5 terraform: update tests 2015-02-19 12:08:06 -08:00
Mitchell Hashimoto
10e82375f2 terraform: early exit and cancellation 2015-02-19 12:08:05 -08:00
Mitchell Hashimoto
d0c77d268a terraform: more tests passing 2015-02-19 12:08:05 -08:00
Mitchell Hashimoto
843cfbe8ad terraform: tainted apply 2015-02-19 12:08:05 -08:00
Mitchell Hashimoto
a3e4b3e966 terraform: more apply tests 2015-02-19 12:08:05 -08:00
Mitchell Hashimoto
691db58478 terraform: apply for orphans 2015-02-19 12:08:05 -08:00
Mitchell Hashimoto
abc68a89a8 terraform: use the proper diff for apply time 2015-02-19 12:08:05 -08:00
Mitchell Hashimoto
aea6b0a7e1 terraform: compare bad diffs for apply 2015-02-19 12:08:05 -08:00
Mitchell Hashimoto
d72ceb597d terraform: more tests 2015-02-19 12:08:05 -08:00
Mitchell Hashimoto
9b7d1509fc terraform: outputs in apply 2015-02-19 12:08:05 -08:00
Mitchell Hashimoto
a614f6105a terraform: apply starting 2015-02-19 12:08:04 -08:00
Mitchell Hashimoto
93f3050dbd terraform: make things more linear 2015-02-19 12:08:04 -08:00
Mitchell Hashimoto
67e7aeeea0 terraform: PLAN IS DONE
http://i.imgur.com/fiw5hXx.gif
2015-02-19 12:08:04 -08:00
Mitchell Hashimoto
f85baf6e26 terraform: more tests 2015-02-19 12:08:04 -08:00