This PR changes the behavior of 'nomad job validate' to forward the
request to the nomad leader, rather than responding from any server.
This is because we need the leader when validating Vault tokens, since
the leader is the only server with an active vault client.
The QEMU driver can take an optional `graceful_shutdown` configuration
which will create a Unix socket to send ACPI shutdown signal to the VM.
Unix sockets have a hard length limit and the driver implementation
assumed that QEMU versions 2.10.1 were able to handle longer paths. This
is not correct, the linked QEMU fix only changed the behaviour from
silently truncating longer socket paths to throwing an error.
By validating the socket path before starting the QEMU machine we can
provide users a more actionable and meaningful error message, and by
using a shorter socket file name we leave a bit more room for
user-defined values in the path, such as the task name.
The maximum length allowed is also platform-dependant, so validation
needs to be different for each OS.
When a QEMU task is recovered the monitor socket path was not being
restored into the task handler, so the `graceful_shutdown` configuration
was effectively ignored if the client restarted.
* Allow specification of CSI staging and publishing directory path
* Add website documentation for stage_publish_dir
* Replace erroneous reference to csi_plugin.mount_config with csi_plugin.mount_dir
* Avoid requiring CSI plugins to be redeployed after introducing StagePublishDir
When we delete a namespace, we check to ensure that there are no non-terminal
jobs, which effectively covers evals, allocs, etc. CSI volumes are also
namespaced, so extend this check to cover CSI volumes.
Return 429 response on HTTP max connection limit. Instead of silently closing
the connection, return a `429 Too Many Requests` HTTP response with a helpful
error message to aid debugging when the connection limit is unintentionally
reached.
Set a 10-millisecond write timeout and rate limiter for connection-limit 429
response to prevent writing the HTTP response from consuming too many server
resources.
Add `nomad.agent.http.exceeded metric` counting the number of HTTP connections
exceeding concurrency limit.
When applying a raft log to expire ACL tokens, we need to use a
timestamp provided by the leader so that the result is deterministic
across servers. Use leader's timestamp from RPC call
Plan rejections occur when the scheduler work and the leader plan
applier disagree on the feasibility of a plan. This may happen for valid
reasons: since Nomad does parallel scheduling, it is expected that
different workers will have a different state when computing placements.
As the final plan reaches the leader plan applier, it may no longer be
valid due to a concurrent scheduling taking up intended resources. In
these situations the plan applier will notify the worker that the plan
was rejected and that they should refresh their state before trying
again.
In some rare and unexpected circumstances it has been observed that
workers will repeatedly submit the same plan, even if they are always
rejected.
While the root cause is still unknown this mitigation has been put in
place. The plan applier will now track the history of plan rejections
per client and include in the plan result a list of node IDs that should
be set as ineligible if the number of rejections in a given time window
crosses a certain threshold. The window size and threshold value can be
adjusted in the server configuration.
To avoid marking several nodes as ineligible at one, the operation is rate
limited to 5 nodes every 30min, with an initial burst of 10 operations.
Fixes#13505
This fixes#13505 by treating reserved_ports like we treat a lot of jobspec settings: merging settings from more global stanzas (client.reserved.reserved_ports) "down" into more specific stanzas (client.host_networks[].reserved_ports).
As discussed in #13505 there are other options, and since it's totally broken right now we have some flexibility:
Treat overlapping reserved_ports on addresses as invalid and refuse to start agents. However, I'm not sure there's a cohesive model we want to publish right now since so much 0.9-0.12 compat code still exists! We would have to explain to folks that if their -network-interface and host_network addresses overlapped, they could only specify reserved_ports in one place or the other?! It gets ugly.
Use the global client.reserved.reserved_ports value as the default and treat host_network[].reserverd_ports as overrides. My first suggestion in the issue, but @groggemans made me realize the addresses on the agent's interface (as configured by -network-interface) may overlap with host_networks, so you'd need to remove the global reserved_ports from addresses shared with a shared network?! This seemed really confusing and subtle for users to me.
So I think "merging down" creates the most expressive yet understandable approach. I've played around with it a bit, and it doesn't seem too surprising. The only frustrating part is how difficult it is to observe the available addresses and ports on a node! However that's a job for another PR.
* Failing test and TODO for wildcard
* Alias the namespace query parameter for Evals
* eval: fix list when using ACLs and * namespace
Apply the same verification process as in job, allocs and scaling
policy list endpoints to handle the eval list when using an ACL token
with limited namespace support but querying using the `*` wildcard
namespace.
* changelog: add entry for #13530
* ui: set namespace when querying eval
Evals have a unique UUID as ID, but when querying them the Nomad API
still expects a namespace query param, otherwise it assumes `default`.
Co-authored-by: Luiz Aoqui <luiz@hashicorp.com>
Stream snapshot to FSM when restoring from archive
The `RestoreFromArchive` helper decompresses the snapshot archive to a
temporary file before reading it into the FSM. For large snapshots
this performs a lot of disk IO. Stream decompress the snapshot as we
read it, without first writing to a temporary file.
Add bexpr filters to the `RestoreFromArchive` helper.
The operator can pass these as `-filter` arguments to `nomad operator
snapshot state` (and other commands in the future) to include only
desired data when reading the snapshot.
Use the same output format when listing multiple evals in the `eval
list` command and when `eval status <prefix>` matches more than one
eval.
Include the eval namespace in all output formats and always include the
job ID in `eval status` since, even `node-update` evals are related to a
job.
Add Node ID to the evals table output to help differentiate
`node-update` evals.
Co-authored-by: James Rasell <jrasell@hashicorp.com>
Whenever a node joins the cluster, either for the first time or after
being `down`, we emit a evaluation for every system job to ensure all
applicable system jobs are running on the node.
This patch adds an optimization to skip creating evaluations for system
jobs not in the current node's DC. While the scheduler performs the same
feasability check, skipping the creation of the evaluation altogether
saves disk, network, and memory.
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.
api: apply new ACL check for wildcard namespace
In #13606 the ACL check was refactored to better support the all
namespaces wildcard (`*`). This commit applies the changes to the jobs
and alloc list endpoints.
The `operator debug` command doesn't output the leader anywhere in the
output, which adds extra burden to offline debugging (away from an
ongoing incident where you can simply check manually). Query the
`/v1/status/leader` API but degrade gracefully.
* core: allow pause/un-pause of eval broker on region leader.
* agent: add ability to pause eval broker via scheduler config.
* cli: add operator scheduler commands to interact with config.
* api: add ability to pause eval broker via scheduler config
* e2e: add operator scheduler test for eval broker pause.
* docs: include new opertor scheduler CLI and pause eval API info.
When calculating a services namespace for registration, the code
assumed the first task within the task array would include a
service block. This is incorrect as it is possible only a latter
task within the array contains a service definition.
This change fixes the logic, so we correctly search for a service
definition before identifying the namespace.
This PR adds the 'choose' query parameter to the '/v1/service/<service>' endpoint.
The value of 'choose' is in the form '<number>|<key>', number is the number
of desired services and key is a value unique but consistent to the requester
(e.g. allocID).
Folks aren't really expected to use this API directly, but rather through consul-template
which will soon be getting a new helper function making use of this query parameter.
Example,
curl 'localhost:4646/v1/service/redis?choose=2|abc123'
Note: consul-templte v0.29.1 includes the necessary nomadServices functionality.
The plan applier has to get a snapshot with a minimum index for the
plan it's working on in order to ensure consistency. Under heavy raft
loads, we can exceed the timeout. When this happens, we hit a bug
where the plan applier blocks waiting on the `indexCh` forever, and
all schedulers will block in `Plan.Submit`.
Closing the `indexCh` when the `asyncPlanWait` is done with it will
prevent the deadlock without impacting correctness of the previous
snapshot index.
This changeset includes the a PoC failing test that works by injecting
a large timeout into the state store. We need to turn this into a test
we can run normally without breaking the state store before we can
merge this PR.
Increase `snapshotMinIndex` timeout to 10s.
This timeout creates backpressure where any concurrent `Plan.Submit`
RPCs will block waiting for results. This sheds load across all
servers and gives raft some CPU to catch up, because schedulers won't
dequeue more work while waiting. Increase it to 10s based on
observations of large production clusters.
The changelog entry for #13340 indicated it was an improvement. But on
discussion, it was determined that this was a workaround for a
regression. Update the changelog to make this clear.
If the node has been GC'd or is down, we can't send it a node
unpublish. The CSI spec requires that we don't send the controller
unpublish before the node unpublish, but in the case where a node is
gone we can't know the final fate of the node unpublish step.
The `csi_hook` on the client will unpublish if the allocation has
stopped and if the host is terminated there's no mount for the volume
anyways. So we'll now assume that the node has unpublished at its
end. If it hasn't, any controller unpublish will potentially hang or
error and need to be retried.
As a performance optimization in the scheduler, feasibility checks
that apply to an entire class are only checked once for all nodes of
that class. Other feasibility checks are "available" checks because
they rely on more ephemeral characteristics and don't contribute to
the hash for the node class. This currently includes only CSI.
We have a separate fast path for "available" checks when the node has
already been marked eligible on the basis of class. This fast path has
a bug where it returns early rather than continuing the loop. This
causes the entire task group to be rejected.
Fix the bug by not returning early in the fast path and instead jump
to the top of the loop like all the other code paths in this method.
Includes a new test exercising topology at whole-scheduler level and a
fix for an existing test that should've caught this previously.
This PR update to Go 1.18.2. Also update the versions of hclfmt
and go-hclogfmt which includes newer dependencies necessary for dealing
with go1.18.
The hcl v2 branch is now 'nomad-v2.9.1+tweaks2', to include a fix for
newer macOS versions: 8927e75e82
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.
* LastIndexOf and always append a namespace on job links
* Confirmed the volume equivalent and simplified idWIthNamespace logic
* Changelog added
* PR comments addressed
* Drop the redirect for the time being
* Tests updated to reflect namespace on links
* Task detail test default namespace link for test
* chore: run prettier on hbs files
* ui: ensure to pass a real job object to task-group link
* chore: add changelog entry
* chore: prettify template
* ui: template helper for formatting jobId in LinkTo component
* ui: handle async relationship
* ui: pass in job id to model arg instead of job model
* update test for serialized namespace
* ui: defend against null in tests
* ui: prettified template added whitespace
* ui: rollback ember-data to 3.24 because watcher return undefined on abort
* ui: use format-job-helper instead of job model via alloc
* ui: fix whitespace in template caused by prettier using template helper
* ui: update test for new namespace
* ui: revert prettier change
Co-authored-by: Luiz Aoqui <luiz@hashicorp.com>
This test exercises upgrades between 0.8 and Nomad versions greater
than 0.9. We have not supported 0.8.x in a very long time and in any
case the test has been marked to skip because the downloader doesn't
work.
Fixes#10200
**The bug**
A user reported receiving the following error when an alloc was placed
that needed to preempt existing allocs:
```
[ERROR] client.alloc_watcher: error querying previous alloc:
alloc_id=28... previous_alloc=8e... error="rpc error: alloc lookup
failed: index error: UUID must be 36 characters"
```
The previous alloc (8e) was already complete on the client. This is
possible if an alloc stops *after* the scheduling decision was made to
preempt it, but *before* the node running both allocations was able to
pull and start the preemptor. While that is hopefully a narrow window of
time, you can expect it to occur in high throughput batch scheduling
heavy systems.
However the RPC error made no sense! `previous_alloc` in the logs was a
valid 36 character UUID!
**The fix**
The fix is:
```
- prevAllocID: c.Alloc.PreviousAllocation,
+ prevAllocID: watchedAllocID,
```
The alloc watcher new func used for preemption improperly referenced
Alloc.PreviousAllocation instead of the passed in watchedAllocID. When
multiple allocs are preempted, a watcher is created for each with
watchedAllocID set properly by the caller. In this case
Alloc.PreviousAllocation="" -- which is where the `UUID must be 36 characters`
error was coming from! Sadly we were properly referencing
watchedAllocID in the log, so it made the error make no sense!
**The repro**
I was able to reproduce this with a dev agent with [preemption enabled](https://gist.github.com/schmichael/53f79cbd898afdfab76865ad8c7fc6a0#file-preempt-hcl)
and [lowered limits](https://gist.github.com/schmichael/53f79cbd898afdfab76865ad8c7fc6a0#file-limits-hcl)
for ease of repro.
First I started a [low priority count 3 job](https://gist.github.com/schmichael/53f79cbd898afdfab76865ad8c7fc6a0#file-preempt-lo-nomad),
then a [high priority job](https://gist.github.com/schmichael/53f79cbd898afdfab76865ad8c7fc6a0#file-preempt-hi-nomad)
that evicts 2 low priority jobs. Everything worked as expected.
However if I force it to use the [remotePrevAlloc implementation](https://github.com/hashicorp/nomad/blob/v1.3.0-beta.1/client/allocwatcher/alloc_watcher.go#L147),
it reproduces the bug because the watcher references PreviousAllocation
instead of watchedAllocID.
The task runner hook `Prestart` response object includes a `Done`
field that's intended to tell the client not to run the hook
again. The plugin supervisor creates mount points for the task during
prestart and saves these mounts in the hook resources. But if a client
restarts the hook resources will not be populated. If the plugin task
restarts at any time after the client restarts, it will fail to have
the correct mounts and crash loop until restart attempts run out.
Fix this by not returning `Done` in the response, just as we do for
the `volume_mount_hook`.
