We use capped exponential backoff in several places in the code when handling
failures. The code we've copy-and-pasted all over has a check to see if the
backoff is greater than the limit, but this check happens after the bitshift and
we always increment the number of attempts. This causes an overflow with a
fairly small number of failures (ex. at one place I tested it occurs after only
24 iterations), resulting in a negative backoff which then never recovers. The
backoff becomes a tight loop consuming resources and/or DoS'ing a Nomad RPC
handler or an external API such as Vault. Note this doesn't occur in places
where we cap the number of iterations so the loop breaks (usually to return an
error), so long as the number of iterations is reasonable.
Introduce a helper with a check on the cap before the bitshift to avoid overflow in all
places this can occur.
Fixes: #18199
Co-authored-by: stswidwinski <stan.swidwinski@gmail.com>
* Throw away result of multierror.Append
When given a *multierror.Error, it is mutated, therefore the return
value is not needed.
* Simplify MergeMultierrorWarnings, use StringBuilder
* Hash.Write() never returns an error
* Remove error that was always nil
* Remove error from Resources.Add signature
When this was originally written it could return an error, but that was
refactored away, and callers of it as of today never handle the error.
* Throw away results of io.Copy during Bridge
* Handle errors when computing node class in test
Even if a plugin sends back an empty `[]*device.DeviceGroup`, it's transformed to `nil` during the RPC. Our custom device plugin is returning empty `FingerprintResponse.Devices` very often. Our temporary fix is to send a dummy `*DeviceGroup` if the slice is empty. This has the effect of never triggering the "first fingerprint" and therefore timing out after 50s.
In turn, this made our node exceed its hearbeat grace period when restarting it, revoking all vault tokens for its allocations, causing a restart of all our allocations because the token couldn't be renewed.
Removing the logic for `f.Devices == nil` does not appear to affect the functionality of the function.
Fix some docstring typos and fix noisy log message during client restarts.
A log for the common case where the plugin socket isn't ready yet
isn't actionable by the operator so having it at info is just noise.
Fixes a bug where we cpu is pigged at 100% due to collecting devices
statistics. The passed stats interval was ignored, and the default zero
value causes a very tight loop of stats collection.
FWIW, in my testing, it took 2.5-3ms to collect nvidia GPU stats, on a
`g2.2xlarge` ec2 instance.
The stats interval defaults to 1 second and is user configurable. I
believe this is too frequent as a default, and I may advocate for
reducing it to a value closer to 5s or 10s, but keeping it as is for
now.
Fixes https://github.com/hashicorp/nomad/issues/6057 .
The driver manager is modeled after the device manager and is started by the client.
It's responsible for handling driver lifecycle and reattachment state, as well as
processing the incomming fingerprint and task events from each driver. The mananger
exposes a method for registering event handlers for task events that is used by the
task runner to update the server when a task has been updated with an event.
Since driver fingerprinting has been implemented by the driver manager, it is no
longer needed in the fingerprint mananger and has been removed.
This PR introduces a device hook that retrieves the device mount
information for an allocation. It also updates the computed node class
computation to take into account devices.
TODO Fix the task runner unit test. The environment variable is being
lost even though it is being properly set in the prestart hook.
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.
Introduce a device manager that manages the lifecycle of device plugins
on the client. It fingerprints, collects stats, and forwards Reserve
requests to the correct plugin. The manager, also handles device plugins
failing and validates their output.
