This change makes few compromises:
* Looks up the devices associated with tasks at look up time. Given
that `nomad alloc status` is called rarely generally (compared to stats
telemetry and general job reporting), it seems fine. However, the
lookup overhead grows bounded by number of `tasks x total-host-devices`,
which can be significant.
* `client.Client` performs the task devices->statistics lookup. It
passes self to alloc/task runners so they can look up the device statistics
allocated to them.
* Currently alloc/task runners are responsible for constructing the
entire RPC response for stats
* The alternatives for making task runners device statistics aware
don't seem appealing (e.g. having task runners contain reference to hostStats)
* On the alloc aggregation resource usage, I did a naive merging of task device statistics.
* Personally, I question the value of such aggregation, compared to
costs of struct duplication and bloating the response - but opted to be
consistent in the API.
* With naive concatination, device instances from a single device group used by separate tasks in the alloc, would be aggregated in two separate device group statistics.
This adds constraints for asserting that a given attribute or value
exists, or does not exist. This acts as a companion to =, or !=
operators, e.g:
```hcl
constraint {
attribute = "${attrs.type}"
operator = "!="
value = "database"
}
constraint {
attribute = "${attrs.type}"
operator = "is_set"
}
```
In state values, we need to be able to distinguish between zero values
(e.g. `false`) and unset values (e.g. `nil`).
We can alternatively use protobuf `oneOf` and nested map to ensure
consistency of fields that are set together, but the golang
representation does not represent that well and introducing a mismatch
between representations. Thus, I opted not to use it.
* upgrade Nomad from 0.8.4 to 0.8.6
* update deprecated nomad and vault commands
* update AMI ID
* add ingress rule for default fabio port and fabio UI
* upgrade Consul and Vault versions
* update AMI ID in README.md and terraform.tfvars
This commit allows the ConstraintChecker to test values that do not exist.
This is useful when wanting to _exclude_ given nodes from executing a
job, for example, if you wanted to give canary nodes an attribute, and
not run critical services on them, you may specify something like the
below, but not want to tag all other nodes with the inverse.
```hcl
constraint {
attribute = "${node.attr.canary}
operator = "!="
value = "1"
}
```
This also requires all constraint checkers to allow for nil target
values, as they will no longer be short circuited by resolving a target.
We run with ~120 apt sources, and apt-get update fails if any of them is
down.
True errors would be raised again at install phase as true dependencies
fetch would fail.
;
nomad/client take very long and exceed 15m sometimes:
In https://travis-ci.org/hashicorp/nomad/jobs/452990197 :
```
panic: test timed out after 15m0s
goroutine 4739 [running]:
testing.(*M).startAlarm.func1()
/home/travis/.gimme/versions/go1.11.2.linux.amd64/src/testing/testing.go:1296 +0xfd
....
goroutine 4665 [select]:
github.com/hashicorp/nomad/vendor/google.golang.org/grpc.newClientStream.func5(0xc0003dd500, 0xc000420120, 0x2b3f86295588, 0xc000496810)
/home/travis/gopath/src/github.com/hashicorp/nomad/vendor/google.golang.org/grpc/stream.go:287 +0xd7
created by github.com/hashicorp/nomad/vendor/google.golang.org/grpc.newClientStream
/home/travis/gopath/src/github.com/hashicorp/nomad/vendor/google.golang.org/grpc/stream.go:286 +0x842
FAIL github.com/hashicorp/nomad/client/driver 900.036s
```
Tests expect that as soon as eventer shuts down immediately on context
cancellations; but golang does not guarantee priority when multiple
pending channels are ready in a select statement.
The default job here contains some exec task config (for setting
command and args) that aren't used for mock driver. Now, the alloc
runner seems stricter about validating fields and errors on unexpected
fields.
Updating configs in tests so we can have an explicit task config
whenever driver is set explicitly.
This test expects 11 repeats of the same message emitted at intervals of
200ms; so we need more than 2 seconds to adjust for time sleep
variations and the like. So raising it to 3s here that should be
enough.
Mock driver config uses `time.Duration` fields but we initialize them
inconsistently, as time.Duration sometimes and as duration strings other
times. Previously, `mapstructure` handles it and does the right thing.
This is no longer the case with MsgPack. I could not find a good way to
bring back old behavior without too much complexity. `MsgPack` extended
types weren't ideal here as we lose type information (e.g. int64 vs
string), and the input is a generic map and not a MsgPack serialization
of duration.
As such, I went with the simple solution of declaring the config field
as duration string, and panicing if the test doesn't pass a valid
string.
I found this to cause the smallest change in tests, but we can
alternatively force all to be int64 instead.
