The `killTasks` function will kill all the alloc runners
task runners. If the task of a task runner has already
completed, the killing of the task runner can cause
confusion due to the task event showing that the task
was signaled even though it is already complete.
To prevent this, a check is done when creating the
task event to determine if the task has completed. If
it has no task event is created and when the task
runner is killed, no extra task event is added.
Batch job allocations that are drained from a node will be moved
to an eligible node. However, when no eligible nodes are available
to place the draining allocations, the tasks will end up being
complete and will not be placed when an eligible node becomes
available. This occurs because the drained allocations are
simultaneously stopped on the draining node while attempting to
be placed on an eligible node. The stopping of the allocations on
the draining node result in tasks being killed, but importantly this
kill does not fail the task. The result is tasks reporting as complete
due to their state being dead and not being failed. As such, when an
eligible node becomes available, all tasks will show as complete and
no allocations will need to be placed.
To prevent the behavior described above a check is performed when
the alloc runner kills its tasks. If the allocation's job type is
batch, and the allocation has a desired transition of migrate, the
task will be failed when it is killed. This ensures the task does
not report as complete, and when an eligible node becomes available
the allocations are placed as expected.
Some of our allocrunner hooks require a task environment for interpolating values based on the node or allocation. But several of the hooks accept an already-built environment or builder and then keep that in memory. Both of these retain a copy of all the node attributes and allocation metadata, which balloons memory usage until the allocation is GC'd.
While we'd like to look into ways to avoid keeping the allocrunner around entirely (see #25372), for now we can significantly reduce memory usage by creating the task environment on-demand when calling allocrunner methods, rather than persisting it in the allocrunner hooks.
In doing so, we uncover two other bugs:
* The WID manager, the group service hook, and the checks hook have to interpolate services for specific tasks. They mutated a taskenv builder to do so, but each time they mutate the builder, they write to the same environment map. When a group has multiple tasks, it's possible for one task to set an environment variable that would then be interpolated in the service definition for another task if that task did not have that environment variable. Only the service definition interpolation is impacted. This does not leak env vars across running tasks, as each taskrunner has its own builder.
To fix this, we move the `UpdateTask` method off the builder and onto the taskenv as the `WithTask` method. This makes a shallow copy of the taskenv with a deep clone of the environment map used for interpolation, and then overwrites the environment from the task.
* The checks hook interpolates Nomad native service checks only on `Prerun` and not on `Update`. This could cause unexpected deregistration and registration of checks during in-place updates. To fix this, we make sure we interpolate in the `Update` method.
I also bumped into an incorrectly implemented interface in the CSI hook. I've pulled that and some better guardrails out to https://github.com/hashicorp/nomad/pull/25472.
Fixes: https://github.com/hashicorp/nomad/issues/25269
Fixes: https://hashicorp.atlassian.net/browse/NET-12310
Ref: https://github.com/hashicorp/nomad/issues/25372
* Upgrade to using hashicorp/go-metrics@v0.5.4
This also requires bumping the dependencies for:
* memberlist
* serf
* raft
* raft-boltdb
* (and indirectly hashicorp/mdns due to the memberlist or serf update)
Unlike some other HashiCorp products, Nomads root module is currently expected to be consumed by others. This means that it needs to be treated more like our libraries and upgrade to hashicorp/go-metrics by utilizing its compat packages. This allows those importing the root module to control the metrics module used via build tags.
The Nomad client can now optionally emit telemetry data from the
prerun and prestart hooks. This allows operators to monitor and
alert on failures and time taken to complete.
The new datapoints are:
- nomad.client.alloc_hook.prerun.success (counter)
- nomad.client.alloc_hook.prerun.failed (counter)
- nomad.client.alloc_hook.prerun.elapsed (sample)
- nomad.client.task_hook.prestart.success (counter)
- nomad.client.task_hook.prestart.failed (counter)
- nomad.client.task_hook.prestart.elapsed (sample)
The hook execution time is useful to Nomad engineering and will
help optimize code where possible and understand job specification
impacts on hook performance.
Currently only the PreRun and PreStart hooks have telemetry
enabled, so we limit the number of new metrics being produced.
On supported platforms, the secrets directory is a 1MiB tmpfs. But some tasks
need larger space for downloading large secrets. This is especially the case for
tasks using `templates`, which need extra room to write a temporary file to the
secrets directory that gets renamed to the old file atomically.
This changeset allows increasing the size of the tmpfs in the `resources`
block. Because this is a memory resource, we need to include it in the memory we
allocate for scheduling purposes. The task is already prevented from using more
memory in the tmpfs than the `resources.memory` field allows, but can bypass
that limit by writing to the tmpfs via `template` or `artifact` blocks.
Therefore, we need to account for the size of the tmpfs in the allocation
resources. Simply adding it to the memory needed when we create the allocation
allows it to be accounted for in all downstream consumers, and then we'll
subtract that amount from the memory resources just before configuring the task
driver.
For backwards compatibility, the default value of 1MiB is "free" and ignored by
the scheduler. Otherwise we'd be increasing the allocated resources for every
existing alloc, which could cause problems across upgrades. If a user explicitly
sets `resources.secrets = 1` it will no longer be free.
Fixes: https://github.com/hashicorp/nomad/issues/2481
Ref: https://hashicorp.atlassian.net/browse/NET-10070
This change exposes CNI configuration details of a network
namespace as environment variables. This allows a task to use
these value to configure itself; a potential use case is to run
a Raft application binding to IP and Port details configured using
the bridge network mode.
Services can have some of their string fields interpolated. The new Workload
Identity flow doesn't interpolate the services before requesting signed
identities or using those identities to get Consul tokens.
Add support for interpolation to the WID manager and the Consul tokens hook by
providing both with a taskenv builder. Add an "interpolate workload" field to
the WI handle to allow passing the original workload name to the server so the
server can find the correct service to sign.
This changeset also makes two related test improvements:
* Remove the mock WID manager, which was only used in the Consul hook tests and
isn't necessary so long as we provide the real WID manager with the mock
signer and never call `Run` on it. It wasn't feasible to exercise the correct
behavior without this refactor, as the mocks were bypassing the new code.
* Fixed swapped expect-vs-actual assertions on the `consul_hook` tests.
Fixes: https://github.com/hashicorp/nomad/issues/20025
* exec2: add client support for unveil filesystem isolation mode
This PR adds support for a new filesystem isolation mode, "Unveil". The
mode introduces a "alloc_mounts" directory where tasks have user-owned
directory structure which are bind mounts into the real alloc directory
structure. This enables a task driver to use landlock (and maybe the
real unveil on openbsd one day) to isolate a task to the task owned
directory structure, providing sandboxing.
* actually create alloc-mounts-dir directory
* fix doc strings about alloc mount dir paths
* exec2: implement dynamic workload users taskrunner hook
This PR impelements a TR hook for allocating dynamic workload users from
a pool managed by the Nomad client. This adds a new task driver Capability,
DynamicWorkloadUsers - which a task driver must indicate in order to make
use of this feature.
The client config plumbing is coming in a followup PR - in the RFC we
realized having a client.users block would be nice to have, with some
additional unrelated options being moved from the deprecated client.options
config.
* learn to spell
CNI plugins may set DNS configuration, but this isn't threaded through to the
task configuration so that we can write it to the `/etc/resolv.conf` file as
needed. Add the `AllocNetworkStatus` to the alloc hook resources so they're
accessible from the taskrunner. Any DNS entries provided by the user will
override these values.
Fixes: https://github.com/hashicorp/nomad/issues/11102
* client/allocdir: use an interface in place of AllocDir structs
This PR replace *allocdir.AllocDir with allocdir.Interface such that we
may eventually have another implementation of alloc directories. This is
in support of the exec2 driver, which will need an implementation of the
alloc directory incompatibile with the current version.
* use rlock
The allocrunner has a service registration handler that proxies various API
calls to Consul. With multi-cluster support (for ENT), the service registration
handler is what selects the correct Consul client. The name of this field in the
allocrunner and taskrunner code base looks like it's referring to the actual
Consul API client. This was actually the case before Nomad native service
discovery was implemented, but now the name is misleading.
When agents start, they create a shared Consul client that is then wrapped as
various interfaces for testability, and used in constructing the Nomad client
and server. The interfaces that support workload services (rather than the Nomad
agent itself) need to support multiple Consul clusters for Nomad
Enterprise. Update these interfaces to be factory functions that return the
Consul client for a given cluster name. Update the `ServiceClient` to split
workload updates between clusters by creating a wrapper around all the clients
that delegates to the cluster-specific `ServiceClient`.
Ref: https://github.com/hashicorp/team-nomad/issues/404
When clients are restarted and the identity hook runs when we restore
allocations, the running allocations are likely to have already-signed Workload
Identities that are unexpired. Save these to the client's local state DB so that
we can avoid a thundering herd of RPCs during client restart. When we restore,
we'll check if there's at least one expired signed WI before making any initial
signing request.
Included:
* Renames `getIdentities` to `getInitialIdentities` to make the workflow more clear.
* Renames the existing `widmgr_test.go` file of integration tests, which is in its
own package to avoid circular imports to `widmgr_int_test.go`
This commit splits identity_hook between the allocrunner and taskrunner. The
allocrunner-level part of the hook signs each task identity, and the
taskrunner-level part picks it up and stores secrets for each task.
The code revamps the WIDMgr, which is now split into 2 interfaces:
IdentityManager which manages renewals of signatures and handles sending
updates to subscribers via Watch method, and IdentitySigner which only does the
signing.
This work is necessary for having a unified Consul login workflow that comes
with the new Consul integration. A new, allocrunner-level consul_hook will now
be the only hook doing Consul authentication.
Nomad Enterprise will support configuring multiple Vault clients. Instead of
having a single Vault client field in the Nomad client, we'll have a function
that callers can parameterize by the Vault cluster name that returns the
correctly configured Vault API client wrapper.
* client: refactor cpuset partitioning
This PR updates the way Nomad client manages the split between tasks
that make use of resources.cpus vs. resources.cores.
Previously, each task was explicitly assigned which CPU cores they were
able to run on. Every time a task was started or destroyed, all other
tasks' cpusets would need to be updated. This was inefficient and would
crush the Linux kernel when a client would try to run ~400 or so tasks.
Now, we make use of cgroup heirarchy and cpuset inheritence to efficiently
manage cpusets.
* cr: tweaks for feedback
Allows for multiple `identity{}` blocks for tasks along with user-specified audiences. This is a building block to allow workload identities to be used with Consul, Vault and 3rd party JWT based auth methods.
Expiration is still unimplemented and is necessary for JWTs to be used securely, so that's up next.
---------
Co-authored-by: Tim Gross <tgross@hashicorp.com>
When claiming a CSI volume, we need to ensure the CSI node plugin is running
before we send any CSI RPCs. This extends even to the controller publish RPC
because it requires the storage provider's "external node ID" for the
client. This primarily impacts client restarts but also is a problem if the node
plugin exits (and fingerprints) while the allocation that needs a CSI volume
claim is being placed.
Unfortunately there's no mapping of volume to plugin ID available in the
jobspec, so we don't have enough information to wait on plugins until we either
get the volume from the server or retrieve the plugin ID from data we've
persisted on the client.
If we always require getting the volume from the server before making the claim,
a client restart for disconnected clients will cause all the allocations that
need CSI volumes to fail. Even while connected, checking in with the server to
verify the volume's plugin before trying to make a claim RPC is inherently racy,
so we'll leave that case as-is and it will fail the claim if the node plugin
needed to support a newly-placed allocation is flapping such that the node
fingerprint is changing.
This changeset persists a minimum subset of data about the volume and its plugin
in the client state DB, and retrieves that data during the CSI hook's prerun to
avoid re-claiming and remounting the volume unnecessarily.
This changeset also updates the RPC handler to use the external node ID from the
claim whenever it is available.
Fixes: #13028
In #17354 we made client updates prioritized to reduce client-to-server
traffic. When the client has no previously-acknowledged update we assume that
the update is of typical priority; although we don't know that for sure in
practice an allocation will never become healthy quickly enough that the first
update we send is the update saying the alloc is healthy.
But that doesn't account for allocations that quickly fail in an unrecoverable
way because of allocrunner hook failures, and it'd be nice to be able to send
those failure states to the server more quickly. This changeset does so and adds
some extra comments on reasoning behind priority.
The allocrunner sends several updates to the server during the early lifecycle
of an allocation and its tasks. Clients batch-up allocation updates every 200ms,
but experiments like the C2M challenge has shown that even with this batching,
servers can be overwhelmed with client updates during high volume
deployments. Benchmarking done in #9451 has shown that client updates can easily
represent ~70% of all Nomad Raft traffic.
Each allocation sends many updates during its lifetime, but only those that
change the `ClientStatus` field are critical for progressing a deployment or
kicking off a reschedule to recover from failures.
Add a priority to the client allocation sync and update the `syncTicker`
receiver so that we only send an update if there's a high priority update
waiting, or on every 5th tick. This means when there are no high priority
updates, the client will send updates at most every 1s instead of
200ms. Benchmarks have shown this can reduce overall Raft traffic by 10%, as
well as reduce client-to-server RPC traffic.
This changeset also switches from a channel-based collection of updates to a
shared buffer, so as to split batching from sending and prevent backpressure
onto the allocrunner when the RPC is slow. This doesn't have a major performance
benefit in the benchmarks but makes the implementation of the prioritized update
simpler.
Fixes: #9451
* client: ignore restart issued to terminal allocations
This PR fixes a bug where issuing a restart to a terminal allocation
would cause the allocation to run its hooks anyway. This was particularly
apparent with group_service_hook who would then register services but
then never deregister them - as the allocation would be effectively in
a "zombie" state where it is prepped to run tasks but never will.
* e2e: add e2e test for alloc restart zombies
* cl: tweak text
Co-authored-by: Tim Gross <tgross@hashicorp.com>
---------
Co-authored-by: Tim Gross <tgross@hashicorp.com>
Tools like `nomad-nodesim` are unable to implement a minimal implementation of
an allocrunner so that we can test the client communication without having to
lug around the entire allocrunner/taskrunner code base. The allocrunner was
implemented with an interface specifically for this purpose, but there were
circular imports that made it challenging to use in practice.
Move the AllocRunner interface into an inner package and provide a factory
function type. Provide a minimal test that exercises the new function so that
consumers have some idea of what the minimum implementation required is.
When client nodes are restarted, all allocations that have been scheduled on the
node have their modify index updated, including terminal allocations. There are
several contributing factors:
* The `allocSync` method that updates the servers isn't gated on first contact
with the servers. This means that if a server updates the desired state while
the client is down, the `allocSync` races with the `Node.ClientGetAlloc`
RPC. This will typically result in the client updating the server with "running"
and then immediately thereafter "complete".
* The `allocSync` method unconditionally sends the `Node.UpdateAlloc` RPC even
if it's possible to assert that the server has definitely seen the client
state. The allocrunner may queue-up updates even if we gate sending them. So
then we end up with a race between the allocrunner updating its internal state
to overwrite the previous update and `allocSync` sending the bogus or duplicate
update.
This changeset adds tracking of server-acknowledged state to the
allocrunner. This state gets checked in the `allocSync` before adding the update
to the batch, and updated when `Node.UpdateAlloc` returns successfully. To
implement this we need to be able to equality-check the updates against the last
acknowledged state. We also need to add the last acknowledged state to the
client state DB, otherwise we'd drop unacknowledged updates across restarts.
The client restart test has been expanded to cover a variety of allocation
states, including allocs stopped before shutdown, allocs stopped by the server
while the client is down, and allocs that have been completely GC'd on the
server while the client is down. I've also bench tested scenarios where the task
workload is killed while the client is down, resulting in a failed restore.
Fixes#16381
to avoid leaking task resources (e.g. containers,
iptables) if allocRunner prerun fails during
restore on client restart.
now if prerun fails, TaskRunner.MarkFailedKill()
will only emit an event, mark the task as failed,
and cancel the tr's killCtx, so then ar.runTasks()
-> tr.Run() can take care of the actual cleanup.
removed from (formerly) tr.MarkFailedDead(),
now handled by tr.Run():
* set task state as dead
* save task runner local state
* task stop hooks
also done in tr.Run() now that it's not skipped:
* handleKill() to kill tasks while respecting
their shutdown delay, and retrying as needed
* also includes task preKill hooks
* clearDriverHandle() to destroy the task
and associated resources
* task exited hooks
The allocrunner has a facility for passing data written by allocrunner hooks to
taskrunner hooks. Currently the only consumers of this facility are the
allocrunner CSI hook (which writes data) and the taskrunner volume hook (which
reads that same data).
The allocrunner hook for CSI volumes doesn't set the alloc hook resources
atomically. Instead, it gets the current resources and then writes a new version
back. Because the CSI hook is currently the only writer and all readers happen
long afterwards, this should be safe but #16623 shows there's some sequence of
events during restore where this breaks down.
Refactor hook resources so that hook data is accessed via setters and getters
that hold the mutex.
* client: run alloc pre-kill hooks on last pass despite no live tasks
This PR fixes a bug where alloc pre-kill hooks were not run in the
edge case where there are no live tasks remaining, but it is also
the final update to process for the (terminal) allocation. We need
to run cleanup hooks here, otherwise they will not run until the
allocation gets garbage collected (i.e. via Destroy()), possibly
at a distant time in the future.
Fixes#15477
* client: do not run ar cleanup hooks if client is shutting down
* allocrunner: handle lifecycle when all tasks die
When all tasks die the Coordinator must transition to its terminal
state, coordinatorStatePoststop, to unblock poststop tasks. Since this
could happen at any time (for example, a prestart task dies), all states
must be able to transition to this terminal state.
* allocrunner: implement different alloc restarts
Add a new alloc restart mode where all tasks are restarted, even if they
have already exited. Also unifies the alloc restart logic to use the
implementation that restarts tasks concurrently and ignores
ErrTaskNotRunning errors since those are expected when restarting the
allocation.
* allocrunner: allow tasks to run again
Prevent the task runner Run() method from exiting to allow a dead task
to run again. When the task runner is signaled to restart, the function
will jump back to the MAIN loop and run it again.
The task runner determines if a task needs to run again based on two new
task events that were added to differentiate between a request to
restart a specific task, the tasks that are currently running, or all
tasks that have already run.
* api/cli: add support for all tasks alloc restart
Implement the new -all-tasks alloc restart CLI flag and its API
counterpar, AllTasks. The client endpoint calls the appropriate restart
method from the allocrunner depending on the restart parameters used.
* test: fix tasklifecycle Coordinator test
* allocrunner: kill taskrunners if all tasks are dead
When all non-poststop tasks are dead we need to kill the taskrunners so
we don't leak their goroutines, which are blocked in the alloc restart
loop. This also ensures the allocrunner exits on its own.
* taskrunner: fix tests that waited on WaitCh
Now that "dead" tasks may run again, the taskrunner Run() method will
not return when the task finishes running, so tests must wait for the
task state to be "dead" instead of using the WaitCh, since it won't be
closed until the taskrunner is killed.
* tests: add tests for all tasks alloc restart
* changelog: add entry for #14127
* taskrunner: fix restore logic.
The first implementation of the task runner restore process relied on
server data (`tr.Alloc().TerminalStatus()`) which may not be available
to the client at the time of restore.
It also had the incorrect code path. When restoring a dead task the
driver handle always needs to be clear cleanly using `clearDriverHandle`
otherwise, after exiting the MAIN loop, the task may be killed by
`tr.handleKill`.
The fix is to store the state of the Run() loop in the task runner local
client state: if the task runner ever exits this loop cleanly (not with
a shutdown) it will never be able to run again. So if the Run() loops
starts with this local state flag set, it must exit early.
This local state flag is also being checked on task restart requests. If
the task is "dead" and its Run() loop is not active it will never be
able to run again.
* address code review requests
* apply more code review changes
* taskrunner: add different Restart modes
Using the task event to differentiate between the allocrunner restart
methods proved to be confusing for developers to understand how it all
worked.
So instead of relying on the event type, this commit separated the logic
of restarting an taskRunner into two methods:
- `Restart` will retain the current behaviour and only will only restart
the task if it's currently running.
- `ForceRestart` is the new method where a `dead` task is allowed to
restart if its `Run()` method is still active. Callers will need to
restart the allocRunner taskCoordinator to make sure it will allow the
task to run again.
* minor fixes
The current implementation for the task coordinator unblocks tasks by
performing destructive operations over its internal state (like closing
channels and deleting maps from keys).
This presents a problem in situations where we would like to revert the
state of a task, such as when restarting an allocation with tasks that
have already exited.
With this new implementation the task coordinator behaves more like a
finite state machine where task may be blocked/unblocked multiple times
by performing a state transition.
This initial part of the work only refactors the task coordinator and
is functionally equivalent to the previous implementation. Future work
will build upon this to provide bug fixes and enhancements.
This PR adds support for specifying checks in services registered to
the built-in nomad service provider.
Currently only HTTP and TCP checks are supported, though more types
could be added later.
This PR fixes a bug where client configuration max_kill_timeout was
not being enforced. The feature was introduced in 9f44780 but seems
to have been removed during the major drivers refactoring.
We can make sure the value is enforced by pluming it through the DriverHandler,
which now uses the lesser of the task.killTimeout or client.maxKillTimeout.
Also updates Event.SetKillTimeout to require both the task.killTimeout and
client.maxKillTimeout so that we don't make the mistake of using the wrong
value - as it was being given only the task.killTimeout before.
Fix numerous go-getter security issues:
- Add timeouts to http, git, and hg operations to prevent DoS
- Add size limit to http to prevent resource exhaustion
- Disable following symlinks in both artifacts and `job run`
- Stop performing initial HEAD request to avoid file corruption on
retries and DoS opportunities.
**Approach**
Since Nomad has no ability to differentiate a DoS-via-large-artifact vs
a legitimate workload, all of the new limits are configurable at the
client agent level.
The max size of HTTP downloads is also exposed as a node attribute so
that if some workloads have large artifacts they can specify a high
limit in their jobspecs.
In the future all of this plumbing could be extended to enable/disable
specific getters or artifact downloading entirely on a per-node basis.
This commit performs refactoring to pull out common service
registration objects into a new `client/serviceregistration`
package. This new package will form the base point for all
client specific service registration functionality.
The Consul specific implementation is not moved as it also
includes non-service registration implementations; this reduces
the blast radius of the changes as well.
Some operators use very long group/task `shutdown_delay` settings to
safely drain network connections to their workloads after service
deregistration. But during incident response, they may want to cause
that drain to be skipped so they can quickly shed load.
Provide a `-no-shutdown-delay` flag on the `nomad alloc stop` and
`nomad job stop` commands that bypasses the delay. This sets a new
desired transition state on the affected allocations that the
allocation/task runner will identify during pre-kill on the client.
Note (as documented here) that using this flag will almost always
result in failed inbound network connections for workloads as the
tasks will exit before clients receive updated service discovery
information and won't be gracefully drained.
Fixes#2522
Skip embedding client.alloc_dir when building chroot. If a user
configures a Nomad client agent so that the chroot_env will embed the
client.alloc_dir, Nomad will happily infinitely recurse while building
the chroot until something horrible happens. The best case scenario is
the filesystem's path length limit is hit. The worst case scenario is
disk space is exhausted.
A bad agent configuration will look something like this:
```hcl
data_dir = "/tmp/nomad-badagent"
client {
enabled = true
chroot_env {
# Note that the source matches the data_dir
"/tmp/nomad-badagent" = "/ohno"
# ...
}
}
```
Note that `/ohno/client` (the state_dir) will still be created but not
`/ohno/alloc` (the alloc_dir).
While I cannot think of a good reason why someone would want to embed
Nomad's client (and possibly server) directories in chroots, there
should be no cause for harm. chroots are only built when Nomad runs as
root, and Nomad disables running exec jobs as root by default. Therefore
even if client state is copied into chroots, it will be inaccessible to
tasks.
Skipping the `data_dir` and `{client,server}.state_dir` is possible, but
this PR attempts to implement the minimum viable solution to reduce risk
of unintended side effects or bugs.
When running tests as root in a vm without the fix, the following error
occurs:
```
=== RUN TestAllocDir_SkipAllocDir
alloc_dir_test.go:520:
Error Trace: alloc_dir_test.go:520
Error: Received unexpected error:
Couldn't create destination file /tmp/TestAllocDir_SkipAllocDir1457747331/001/nomad/test/testtask/nomad/test/testtask/.../nomad/test/testtask/secrets/.nomad-mount: open /tmp/TestAllocDir_SkipAllocDir1457747331/001/nomad/test/.../testtask/secrets/.nomad-mount: file name too long
Test: TestAllocDir_SkipAllocDir
--- FAIL: TestAllocDir_SkipAllocDir (22.76s)
```
Also removed unused Copy methods on AllocDir and TaskDir structs.
Thanks to @eveld for not letting me forget about this!
