In this changeset:
* If a Nomad client node is running both a controller and a node
plugin (which is a common case), then if only the controller or the
node is removed, the plugin was not being updated with the correct
counts.
* The existing test for plugin cleanup didn't go back to the state
store, which normally is ok but is complicated in this case by
denormalization which changes the behavior. This commit makes the
test more comprehensive.
* Set "controller required" when plugin has `PUBLISH_READONLY`. All
known controllers that support `PUBLISH_READONLY` also support
`PUBLISH_UNPUBLISH_VOLUME` but we shouldn't assume this.
* Only create plugins when the allocs for those plugins are
healthy. If we allow a plugin to be created for the first time when
the alloc is not healthy, then we'll recreate deleted plugins when
the job's allocs all get marked terminal.
* Terminal plugin alloc updates should cleanup the plugin. The client
fingerprint can't tell if the plugin is unhealthy intentionally (for
the case of updates or job stop). Allocations that are
server-terminal should delete themselves from the plugin and trigger
a plugin self-GC, the same as an unused node.
Several of the CSI `VolumeCapability` methods return pointers, which
we were then comparing to pointers in the request rather than
dereferencing them and comparing their contents.
This changeset does a more fine-grained comparison of the request vs
the capabilities, and adds better error messaging.
Makes it possible to run Linux Containers On Windows with Nomad alongside Windows Containers. Fingerprint prevents only to run Nomad in Windows 10 with Linux Containers
The nil-check here is left-over from an earlier approach that didn't
get merged. It doesn't do anything for us now as we can't ever pass it
`nil` and if we leave it in the `getVolume` call it guards will panic
anyways.
In order to minimize this change while keeping a simple version of the
behavior, we set `lastOk` to the current time less the intial server
connection timeout. If the client starts and never contacts the
server, it will stop all configured tasks after the initial server
connection grace period, on the assumption that we've been out of
touch longer than any configured `stop_after_client_disconnect`.
The more complex state behavior might be justified later, but we
should learn about failure modes first.
- track lastHeartbeat, the client local time of the last successful
heartbeat round trip
- track allocations with `stop_after_client_disconnect` configured
- trigger allocation destroy (which handles cleanup)
- restore heartbeat/killable allocs tracking when allocs are recovered from disk
- on client restart, stop those allocs after a grace period if the
servers are still partioned
This changeset corrects handling of the `ValidationVolumeCapabilities`
response:
* The CSI spec for the `ValidationVolumeCapabilities` requires that
plugins only set the `Confirmed` field if they've validated all
capabilities. The Nomad client improperly assumes that the lack of a
`Confirmed` field should be treated as a failure. This breaks the
Azure and Linode block storage plugins, which don't set this
optional field.
* The CSI spec also requires that the orchestrator check the validation
responses to guard against older versions of a plugin reporting
"valid" for newer fields it doesn't understand.
During MVP development, we reduced the timeout for controller plugins
to avoid long hangs in GC workers. But now that this work has been
moved to the volume watcher, we can restore the original timeout which
is better suited for the characteristic timescales of some cloud
provider APIs and better matches the behavior of k8s.
We should only remove the `ReadAllocs`/`WriteAllocs` values for a
volume after the claim has entered the "ready to free"
state. The volume will eventually be released as expected. But
querying the volume API will show the volume is released before the
controller unpublish has finished and this can cause a race with
starting new jobs.
Test updates are to cover cases where we're dropping claims but not
running through the whole reaping process.
This changeset adds a subsystem to run on the leader, similar to the
deployment watcher or node drainer. The `Watcher` performs a blocking
query on updates to the `CSIVolumes` table and triggers reaping of
volume claims.
This will avoid tying up scheduling workers by immediately sending
volume claim workloads into their own loop, rather than blocking the
scheduling workers in the core GC job doing things like talking to CSI
controllers
The volume watcher is enabled on leader step-up and disabled on leader
step-down.
The volume claim GC mechanism now makes an empty claim RPC for the
volume to trigger an index bump. That in turn unblocks the blocking
query in the volume watcher so it can assess which claims can be
released for a volume.
This would happen because a no connection error happens after the second request fails, but
that's because it's assumed the second request is to a server node. However, if a user clicks
stderr fast enough, the first and second requests are both to the client node. This changes
the logic to check if the request is to the server before deeming log streaming a total failure.
Typically a failover means that the client can't be reached. However, if
the client does eventually return after the timeout period, the log will
stream indefinitely. This fixes that using an API that wasn't broadly
available at the time this was first written.
