Commit graph

2558 commits

Author SHA1 Message Date
Michael Schurter f788974f8a Merge pull request #3288 from simar7/qemu-improvements
qemu: Add bound checks for memory assignment
2017-10-02 14:47:05 -07:00
Simarpreet Singh d801584c46
qemu: Fix lower memory bound to 128M
Signed-off-by: Simarpreet Singh <simar@linux.com>
2017-10-02 14:29:44 -07:00
Simarpreet Singh 10d7d6dab0
gofmt: format qemu.go and qemu_test.go
Signed-off-by: Simarpreet Singh <simar@linux.com>
2017-10-02 13:16:48 -07:00
Michael Schurter a66c53d45a Remove structs import from api
Goes a step further and removes structs import from api's tests as well
by moving GenerateUUID to its own package.
2017-09-29 10:36:08 -07:00
Michael Schurter 77f1fe40e7 Properly autodetect Docker IP in Windows
Our Docker network plugin autodetection code was erroneously treating
Window's default network `nat` as a plugin and defaulting to it instead
of the host.

Fixes #3218
2017-09-27 16:49:23 -07:00
Michael Schurter a8a87af7ed Only build rkt driver on linux
Build stub for non-linux targets
2017-09-27 14:21:45 -07:00
Simarpreet Singh 3d99e71de8
qemu: Add bound checks for memory assignment
Signed-off-by: Simarpreet Singh <simar@linux.com>
2017-09-26 21:07:48 -07:00
Michael Schurter d7229ce6c5 Merge pull request #3256 from dalegaard/master
Enable rkt driver to use address_mode = 'driver'
2017-09-26 18:04:37 -05:00
Alex Dadgar 4173834231 Enable more linters 2017-09-26 15:26:33 -07:00
Lasse Dalegaard 9f584d1114 Ignore rkt network failure if container died early
If the container dies before the network can be read, we now ignore the
error coming out of the network information polling loop. Nomad will
restart the task regardless, so we might be masking the actual error.

The polling loop for the rkt network information, inside the `Start`
method, was getting a bit unwieldy. It's been refactored out so it's not
a seperate function.
2017-09-27 00:15:27 +02:00
Lasse Dalegaard b43ec57c02 Make rkt port mapping test not exit immediately
The rkt port mapping test currently starts redis with --version, which
obviously makes redis exit again almost immediately. This means that the
container exists before the network status can be queried, and so the
test fails.
2017-09-26 23:10:24 +02:00
Lasse Dalegaard 17d155d316 Improve rkt driver network status poll loop
The network status poll loop will now report any networks it ignored, as
well as a no-networks situations.
2017-09-26 21:49:45 +02:00
Lasse Dalegaard bafd32fda0 Refactor rkt network status loop
The network status poll loop for the rkt drivers `Start` method was a
bit messy, and could not display the last encountered error. Here we
clean it up.
2017-09-26 21:27:12 +02:00
Lasse Dalegaard 5e9e2b07bd Small logging fix in rkt/driver 2017-09-26 19:36:13 +02:00
Lasse Dalegaard 3d25fd3b00 Bump minimum rkt version to 1.27.0.
The changes introduces in #3256 require at least rkt 1.27.0 because of
a bug in the JSON output of `rkt status` in previous versions.

Here we upgrade all references to rkt's minimum version, and also make
travis and vagrant use this version when running tests.

Finally we add a CHANGELOG notice.
2017-09-26 19:15:43 +02:00
Lasse Dalegaard f55f2b8f24 Turn rkt network status failure into Start failure
If the rkt driver cannot get the network status, for a task with a
configured port mapping, it will now fail the Start() call and kill the
task instead of simply logging. This matches the Docker behavior.

If no port map is specified, the warnings will be logged but the task
will be allowed to start.
2017-09-26 10:20:57 +02:00
Lasse Dalegaard 55a2e60e1a Test for rkt driver setting DriverNetwork
To test that the rkt driver correctly sets a DriverNetwork, at least
when a port mapping is requested, we amend the
TestRktDriver_PortsMapping test with a small check.
2017-09-26 09:10:50 +02:00
Lasse Dalegaard 2d307d5beb Discard errors from rkt status and cat-manifest
Since we don't actually show these errors anywhere, just discard them
right away.
2017-09-26 09:05:47 +02:00
Chelsea Holland Komlo b26454cf99 Move setGaugeForAllocationStats to emitClientMetrics 2017-09-25 16:05:49 +00:00
Lasse Dalegaard cbcbe0da2e Expose rkt DriverNetwork
Currently the rkt driver does not expose a DriverNetwork instance after
starting the container, which means that address_mode = 'driver' does
not work.

To get the container network information, we can call `rkt status` on
the UUID of the container and grab the container IP from there.

For the port map, we need to grab the pod manifest as it will tell us
which ports the container exposes. We then cross-reference the
configured port name with the container port names, and use that to
create a correct port mapping.

To avoid doing a (bad) reimplementation of the appc schema(which rkt
uses for its manifest) and rkt apis, we pull those in as vendored
dependencies. The versions used are the same ones that rkt use in their
glide dependency configuration for version 1.28.0.
2017-09-21 00:34:22 +02:00
Lasse Dalegaard 7ac599d509 Use rkt prepare + run-prepared instead of run.
The rkt driver currently executes run and asks that the pod UUID is
written to a file that is then polled for changes for up to five
seconds. Many container fetches will take longer than this, so this
method will often not be able to track the pod UUID reliably.

To avoid this problem, rkt allows pods to be first prepared, which will
return their UUID, and then run as a second invocation.

Here we convert the rkt driver's Start method to use this method
instead. This way, the UUID will always be tracked correctly.
2017-09-21 00:17:31 +02:00
Michael Schurter f92ffe5af5 Merge pull request #3105 from hashicorp/f-876-restart-unhealthy
Restart unhealthy tasks
2017-09-17 19:38:32 -07:00
epipho a16c97394f Fix incorrect docker stats 2017-09-16 00:43:03 -04:00
Michael Schurter 67a4a169a9 Name const after what it represents 2017-09-15 14:57:18 -07:00
Michael Schurter 79a7bf3d7c Cleanup and test restart failure code 2017-09-15 14:54:37 -07:00
Michael Schurter 06ca379da0 Add comments 2017-09-15 14:34:36 -07:00
Michael Schurter 4dbaa52aba Fold SetFailure into SetRestartTriggered 2017-09-14 16:48:39 -07:00
Michael Schurter ed77c0944b DRY up restart handling a bit.
All 3 error/failure cases share restart logic, but 2 of them have
special cased conditions.
2017-09-14 16:48:39 -07:00
Michael Schurter 73fb71ca10 RestartDelay isn't needed as checks are re-added on restarts
@dadgar made the excellent observation in #3105 that TaskRunner removes
and re-registers checks on restarts. This means checkWatcher doesn't
need to do *any* internal restart tracking. Individual checks can just
remove themselves and be re-added when the task restarts.
2017-09-14 16:48:39 -07:00
Michael Schurter 06dd86adbd Remove unused lastStart field 2017-09-14 16:47:41 -07:00
Michael Schurter 0447f79288 Removed partially implemented allocLock 2017-09-14 16:47:41 -07:00
Michael Schurter ade29ecbed Improve check watcher logging and add tests
Also expose a mock Consul Agent to allow testing ServiceClient and
checkWatcher from TaskRunner without actually talking to a real Consul.
2017-09-14 16:47:41 -07:00
Michael Schurter a137676358 Add comments and move delay calc to TaskRunner 2017-09-14 16:46:54 -07:00
Michael Schurter 8a87475498 Use existing restart policy infrastructure 2017-09-14 16:46:54 -07:00
Michael Schurter 22690c5f4c Add check watcher for restarting unhealthy tasks 2017-09-14 16:46:54 -07:00
Alex Dadgar d306da846c changelog and feedback 2017-09-14 14:08:58 -07:00
Alex Dadgar 07ed83fdd5 Non-locked accessors to common Node fields
This PR removes locking around commonly accessed node attributes that do
not need to be locked. The locking could cause nodes to TTL as the
heartbeat code path was acquiring a lock that could be held for an
excessively long time. An example of this is when Vault is inaccessible,
since the fingerprint is run with a lock held but the Vault
fingerprinter makes the API calls with a large timeout.

Fixes https://github.com/hashicorp/nomad/issues/2689
2017-09-14 14:08:26 -07:00
Chelsea Komlo 536d38454b Merge pull request #3191 from hashicorp/b-tagged-metrics-panic
Fix panic in emitting tagged allocation metrics
2017-09-11 14:28:50 -04:00
Armon Dadgar d4aed839d2 Merge pull request #3185 from hashicorp/f-acl-reset
Add ability to reset ACL bootstrap process
2017-09-11 10:47:17 -07:00
Armon Dadgar 3d5ecaafff Address @dadgar feedback 2017-09-11 10:30:59 -07:00
Alex Dadgar b3958faa14 Merge pull request #3187 from hashicorp/b-windows-docker
Fix MemorySwappiness on Windows Docker
2017-09-11 09:56:49 -07:00
Alex Dadgar 1cd8f7523f Merge pull request #3184 from hashicorp/b-docker-logging
Fix docker user specified syslogging
2017-09-11 09:31:33 -07:00
Chelsea Holland Komlo 848af92183 fix panic in emitting tagged metrics 2017-09-11 15:32:37 +00:00
Alex Dadgar d3a9463358 Fix MemorySwappiness on Windows Docker
Fixes https://github.com/hashicorp/nomad/issues/3181
2017-09-10 17:46:45 -07:00
Alex Dadgar 3ec7946b3e Fix invalid CPU stats on Windows
This PR fixes an issue introduced in Nomad 0.6.0 due to
https://github.com/shirou/gopsutil/issues/420. The issue arised from the
fact that the Windows stats from gopsutil reports CPUs in
percentages where we expected ticks.
2017-09-10 15:30:48 -07:00
Alex Dadgar 637ae9580a Fix docker user specified syslogging 2017-09-10 14:57:48 -07:00
James Nugent 448145872f client: Guard against "NaN" values from floats
This commit protects against finding `0.NaN` tokens in JSON streams
because of infinity representation on serialization.
2017-09-08 16:21:07 -05:00
Alex Dadgar 31f9e099d9 Merge pull request #3148 from clinta/purge-stopped
Always purge stopped containers
2017-09-05 17:18:05 -07:00
Alex Dadgar 6fdaf38389 Fix repo name passed to docker credential helpers
This PR fixes the server url passed to docker credential helpers and
fixes stderr capture.

Fixes https://github.com/hashicorp/nomad/issues/2957
2017-09-05 16:43:21 -07:00
Alex Dadgar 21564c7c04 Parse Docker mounts correctly (#3163)
* Parse Docker mounts correctly

This PR fixes the parsing of Docker mounts and adds testing to ensure no
regressions.

Fixes https://github.com/hashicorp/nomad/issues/3156

* Review feedback
2017-09-05 14:02:57 -07:00