Here, we ensure that when leader only responds to RPC calls when state
store is up to date. At leadership transition or launch with restored
state, the server local store might not be caught up with latest raft
logs and may return a stale read.
The solution here is to have an RPC consistency read gate, enabled when
`establishLeadership` completes before we respond to RPC calls.
`establishLeadership` is gated by a `raft.Barrier` which ensures that
all prior raft logs have been applied.
Conversely, the gate is disabled when leadership is lost.
This is very much inspired by https://github.com/hashicorp/consul/pull/3154/files
Rename SnapshotAfter to SnapshotMinIndex. The old name was not
technically accurate. SnapshotAtOrAfter is more accurate, but wordy and
still lacks context about what precisely it is at or after (the index).
SnapshotMinIndex was chosen as it describes the action (snapshot), a
constraint (minimum), and the object of the constraint (index).
The previous commit prevented evaluating plans against a state snapshot
which is older than the snapshot at which the plan was created. This is
correct and prevents failures trying to retrieve referenced objects that
may not exist until the plan's snapshot. However, this is insufficient
to guarantee consistency if the following events occur:
1. P1, P2, and P3 are enqueued with snapshot @ 100
2. Leader evaluates and applies Plan P1 with snapshot @ 100
3. Leader evaluates Plan P2 with snapshot+P1 @ 100
4. P1 commits @ 101
4. Leader evaluates applies Plan P3 with snapshot+P2 @ 100
Since only the previous plan is optimistically applied to the state
store, the snapshot used to evaluate a plan may not contain the N-2
plan!
To ensure plans are evaluated and applied serially we must consider all
previous plan's committed indexes when evaluating further plans.
Therefore combined with the last PR, the minimum index at which to
evaluate a plan is:
min(previousPlanResultIndex, plan.SnapshotIndex)
Plan application should use a state snapshot at or after the Raft index
at which the plan was created otherwise it risks being rejected based on
stale data.
This commit adds a Plan.SnapshotIndex which is set by workers when
submitting plan. SnapshotIndex is set to the Raft index of the snapshot
the worker used to generate the plan.
Plan.SnapshotIndex plays a similar role to PlanResult.RefreshIndex.
While RefreshIndex informs workers their StateStore is behind the
leader's, SnapshotIndex is a way to prevent the leader from using a
StateStore behind the worker's.
Plan.SnapshotIndex should be considered the *lower bound* index for
consistently handling plan application.
Plans must also be committed serially, so Plan N+1 should use a state
snapshot containing Plan N. This is guaranteed for plans *after* the
first plan after a leader election.
The Raft barrier on leader election ensures the leader's statestore has
caught up to the log index at which it was elected. This guarantees its
StateStore is at an index > lastPlanIndex.
- updated region in job metadata that gets persisted to nomad datastore
- fixed many unrelated unit tests that used an invalid region value
(they previously passed because hcl wasn't getting picked up and
the job would default to global region)
Enterprise only.
Disable preemption for service and batch jobs by default.
Maintain backward compatibility in a x.y.Z release. Consider switching
the default for new clusters in the future.
Revert plan_apply.go changes from #5411
Since non-Command Raft messages do not update the StateStore index,
SnapshotAfter may unnecessarily block and needlessly fail in idle
clusters where the last Raft message is a non-Command message.
This is trivially reproducible with the dev agent and a job that has 2
tasks, 1 of which fails.
The correct logic would be to SnapshotAfter the previous plan's index to
ensure consistency. New clusters or newly elected leaders will not have
a previous plan, so the index the leader was elected should be used
instead.
Fix a case where `node.StatusUpdatedAt` was manipulated directly in
memory.
This ensures that StatusUpdatedAt is set in raft layer, and ensures that
the field is updated when node drain/eligibility is updated too.
Previous commit could introduce a deadlock if the capacityChangeCh was
full and the receiving side exited before freeing a slot for the sending
side could send. Flush would then block forever waiting to acquire the
lock just to throw the pending update away.
The race is around getting/setting the chan field, not chan operations,
so only lock around getting the chan field.
I assume the mutex was being released before sending on capacityChangeCh
to avoid blocking in the critical section, but:
1. This is race.
2. capacityChangeCh has a *huge* buffer (8096). If it's full things
already seem Very Bad, and a little backpressure seems appropriate.
This fixes a bug in the state store during plan apply. When
denormalizing preempted allocations it incorrectly set the preemptor's
job during the update. This eventually causes a panic downstream in the
client. Added a test assertion that failed before and passes after this fix
* master: (912 commits)
Update redirects.txt
Added redirect for Spark guide link
client: log when server list changes
docs: mention regression in task config validation
fix update to changelog
update CHANGELOG with datacenter config validation https://github.com/hashicorp/nomad/pull/5665
typo: "atleast" -> "at least"
implement nomad exec for rkt
docs: fixed typo
use pty/tty terminology similar to github.com/kr/pty
vendor github.com/kr/pty
drivers: implement streaming exec for executor based drivers
executors: implement streaming exec
executor: scaffolding for executor grpc handling
client: expose allocated memory per task
client improve a comment in updateNetworks
stalebot: Add 'thinking' as an exempt label (#5684)
Added Sparrow link
update links to use new canonical location
Add redirects for restructing done in GH-5667
...
Fixes#1795
Running restored allocations and pulling what allocations to run from
the server happen concurrently. This means that if a client is rebooted,
and has its allocations rescheduled, it may restart the dead allocations
before it contacts the server and determines they should be dead.
This commit makes tasks that fail to reattach on restore wait until the
server is contacted before restarting.
Currently when an evalbroker is disabled, it still recieves delayed
enqueues via log application in the fsm. This causes an ever growing
heap of evaluations that will never be drained, and can cause memory
issues in larger clusters, or when left running for an extended period
of time without a leader election.
This commit prevents the enqueuing of evaluations while we are
disabled, and relies on the leader restoreEvals routine to handle
reconciling state during a leadership transition.
Existing dequeues during an Enabled->Disabled broker state transition are
handled by the enqueueLocked function dropping evals.
Primarily a cleanup commit, however, currently there is a potential race
condition (that I'm not sure we've ever actually hit) during a flapping
SetEnabled/Disabled state where we may never correctly restart the eval
broker, if it was being called from multiple routines.
Our testing so far indicates that ugorji/go/codec maintains backward
compatiblity with the version we are using now, for purposes of Nomad
serialization.
Using latest ugorji/go allows us to get back to using upstream library,
get get the optimizations benefits in RPC paths (including code
generation optimizations).
ugorji/go introduced two significant changes:
* time binary format in debb8e2d2e. Setting `h.BasicHandle.TimeNotBuiltin = true` restores old behavior
* ugorji/go started honoring `json` tag as well:
v1.1.4 is the latest but has a bug in handling RawString that's fixed in
d09a80c1e0
.
`Index` is already included as part of `WriteMeta` embedding.
This is a backward compatible change: Clients never read the field; and
Server refernces to `EmitNodeEventsResponse.Index` would be using the
value in `WriteMeta`, which is consistent with other response structs.
This command will be used to send a signal to either a single task within an
allocation, or all of the tasks if <task-name> is omitted. If the sent signal
terminates the allocation, it will be treated as if the allocation has crashed,
rather than as if it was operator-terminated.
Signal validation is currently handled by the driver itself and nomad
does not attempt to restrict or validate them.
This adds a `nomad alloc stop` command that can be used to stop and
force migrate an allocation to a different node.
This is built on top of the AllocUpdateDesiredTransitionRequest and
explicitly limits the scope of access to that transition to expose it
under the alloc-lifecycle ACL.
The API returns the follow up eval that can be used as part of
monitoring in the CLI or parsed and used in an external tool.
Revert "fingerprint Constraints and Affinities have Equals, as set"
This reverts commit 596f16fb5f1a4a6766a57b3311af806d22382609.
Revert "client tests assert the independent handling of interface and speed"
This reverts commit 7857ac5993a578474d0570819f99b7b6e027de40.
Revert "structs missed applying a style change from the review"
This reverts commit 658916e3274efa438beadc2535f47109d0c2f0f2.
Revert "client, structs comments"
This reverts commit be2838d6baa9d382a5013fa80ea016856f28ade2.
Revert "client fingerprint updateNetworks preserves the network configuration"
This reverts commit fc309cb430e62d8e66267a724f006ae9abe1c63c.
Revert "client_test cleanup comments from review"
This reverts commit bc0bf4efb9114e699bc662f50c8f12319b6b3445.
Revert "client Networks Equals is set equality"
This reverts commit f8d432345b54b1953a4a4c719b9269f845e3e573.
Revert "struct cleanup indentation in RequestedDevice Equals"
This reverts commit f4746411cab328215def6508955b160a53452da3.
Revert "struct Equals checks for identity before value checking"
This reverts commit 0767a4665ed30ab8d9586a59a74db75d51fd9226.
Revert "fix client-test, avoid hardwired platform dependecy on lo0"
This reverts commit e89dbb2ab182b6368507dbcd33c3342223eb0ae7.
Revert "refactor error in client fingerprint to include the offending data"
This reverts commit a7fed726c6e0264d42a58410d840adde780a30f5.
Revert "add client updateNodeResources to merge but preserve manual config"
This reverts commit 84bd433c7e1d030193e054ec23474380ff3b9032.
Revert "refactor struts.RequestedDevice to have its own Equals"
This reverts commit 689782524090e51183474516715aa2f34908b8e6.
Revert "refactor structs.Resource.Networks to have its own Equals"
This reverts commit 49e2e6c77bb3eaa4577772b36c62205061c92fa1.
Revert "refactor structs.Resource.Devices to have its own Equals"
This reverts commit 4ede9226bb971ae42cc203560ed0029897aec2c9.
Revert "add COMPAT(0.10): Remove in 0.10 notes to impl for structs.Resources"
This reverts commit 49fbaace5298d5ccf031eb7ebec93906e1d468b5.
Revert "add structs.Resources Equals"
This reverts commit 8528a2a2a6450e4462a1d02741571b5efcb45f0b.
Revert "test that fingerprint resources are updated, net not clobbered"
This reverts commit 8ee02ddd23bafc87b9fce52b60c6026335bb722d.
This adds a `nomad alloc restart` command and api that allows a job operator
with the alloc-lifecycle acl to perform an in-place restart of a Nomad
allocation, or a given subtask.
Currently when operators need to log onto a machine where an alloc
is running they will need to perform both an alloc/job status
call and then a call to discover the node name from the node list.
This updates both the job status and alloc status output to include
the node name within the information to make operator use easier.
Closes#2359
Cloess #1180
Consider currently dequeued Evaluation's ModifyIndex when determining
its WaitIndex. Normally the Evaluation itself would already be in the
state store snapshot used to determine the WaitIndex. However, since the FSM
applies Raft messages to the state store concurrently with Dequeueing,
it's possible the currently dequeued Evaluation won't yet exist in the
state store snapshot used by JobsForEval.
This can be solved by always considering the current eval's modify index
and using it if it is greater than all of the evals returned by the
state store.
Track the download status of each artifact independently so that if only
one of many artifacts fails to download, completed artifacts aren't
downloaded again.
This seems to fix TestClientAllocations_GarbageCollectAll_Remote being
flaky.
This test confuses me. It joins 2 servers, but then goes out of its way
to make sure the test client only interacts with one. There are not
enough comments for me to figure out the precise assertions this test is
trying to make.
A good old fashioned wait-for-the-client-to-register seems to fix the
flakiness though. The error was that the node could not be found, so
this makes some sense. However, lots of other tests seem to use the same
"wait for node" logic and don't appear to be flaky, so who knows why
waiting fixes this one.
Passes with -race.
Given that the values will rarely change, specially considering that any
changes would be backward incompatible change. As such, it's simpler to
keep syncing manually in the rare occasion and avoid the syncing code
overhead.
nomad/structs is an internal package and imports many libraries (e.g.
raft, codec) that are not relevant to api clients, and may cause
unnecessary dependency pain (e.g. `github.com/ugorji/go/codec`
version is very old now).
Here, we add a code generator that imports the relevant constants from
`nomad/structs`.
I considered using this approach for other structs, but didn't find a
quick viable way to reduce duplication. `nomad/structs` use values as
struct fields (e.g. `string`), while `api` uses value pointer (e.g.
`*string`) instead. Also, sometimes, `api` structs contain deprecated
fields or additional documentation, so simple copy-paste doesn't work.
For these reasons, I opt to keep the status quo.
Race was test only and due to unlocked map access.
Panic was test only and due to checking a field on a struct even when we
knew the struct was nil.
Race output that was fixed:
```
==================
WARNING: DATA RACE
Read at 0x00c000697dd0 by goroutine 768:
runtime.mapaccess2()
/usr/local/go/src/runtime/map.go:439 +0x0
github.com/hashicorp/nomad/nomad.TestLeader_PeriodicDispatcher_Restore_Adds.func8()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader_test.go:402
+0xe6
github.com/hashicorp/nomad/testutil.WaitForResultRetries()
/home/schmichael/go/src/github.com/hashicorp/nomad/testutil/wait.go:30
+0x5a
github.com/hashicorp/nomad/testutil.WaitForResult()
/home/schmichael/go/src/github.com/hashicorp/nomad/testutil/wait.go:22
+0x57
github.com/hashicorp/nomad/nomad.TestLeader_PeriodicDispatcher_Restore_Adds()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader_test.go:401
+0xb53
testing.tRunner()
/usr/local/go/src/testing/testing.go:827 +0x162
Previous write at 0x00c000697dd0 by goroutine 569:
runtime.mapassign()
/usr/local/go/src/runtime/map.go:549 +0x0
github.com/hashicorp/nomad/nomad.(*PeriodicDispatch).Add()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/periodic.go:224
+0x2eb
github.com/hashicorp/nomad/nomad.(*Server).restorePeriodicDispatcher()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:394
+0x29a
github.com/hashicorp/nomad/nomad.(*Server).establishLeadership()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:234
+0x593
github.com/hashicorp/nomad/nomad.(*Server).leaderLoop()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:117
+0x82e
github.com/hashicorp/nomad/nomad.(*Server).monitorLeadership.func1()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:72
+0x6c
Goroutine 768 (running) created at:
testing.(*T).Run()
/usr/local/go/src/testing/testing.go:878 +0x650
testing.runTests.func1()
/usr/local/go/src/testing/testing.go:1119 +0xa8
testing.tRunner()
/usr/local/go/src/testing/testing.go:827 +0x162
testing.runTests()
/usr/local/go/src/testing/testing.go:1117 +0x4ee
testing.(*M).Run()
/usr/local/go/src/testing/testing.go:1034 +0x2ee
main.main()
_testmain.go:1150 +0x221
Goroutine 569 (running) created at:
github.com/hashicorp/nomad/nomad.(*Server).monitorLeadership()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:70
+0x269
==================
```
Similar to previous commits the delayed eval update chan was set and
access from different goroutines causing a race. Passing the chan on the
stack resolves the race.
Race output from `go test -race -run 'Server_RPC$'` in nomad/
```
==================
WARNING: DATA RACE
Write at 0x00c000339150 by goroutine 63:
github.com/hashicorp/nomad/nomad.(*EvalBroker).flush()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/eval_broker.go:708
+0x3dc
github.com/hashicorp/nomad/nomad.(*EvalBroker).SetEnabled()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/eval_broker.go:174
+0xc4
github.com/hashicorp/nomad/nomad.(*Server).revokeLeadership()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:718
+0x1fd
github.com/hashicorp/nomad/nomad.(*Server).leaderLoop()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:122
+0x95d
github.com/hashicorp/nomad/nomad.(*Server).monitorLeadership.func1()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:72
+0x6c
Previous read at 0x00c000339150 by goroutine 73:
github.com/hashicorp/nomad/nomad.(*EvalBroker).runDelayedEvalsWatcher()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/eval_broker.go:771
+0x176
Goroutine 63 (running) created at:
github.com/hashicorp/nomad/nomad.(*Server).monitorLeadership()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:70
+0x269
Goroutine 73 (running) created at:
github.com/hashicorp/nomad/nomad.(*EvalBroker).SetEnabled()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/eval_broker.go:170
+0x173
github.com/hashicorp/nomad/nomad.(*Server).establishLeadership()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:207
+0x355
github.com/hashicorp/nomad/nomad.(*Server).leaderLoop()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:117
+0x82e
github.com/hashicorp/nomad/nomad.(*Server).monitorLeadership.func1()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:72
+0x6c
==================
```
PeriodicDispatch.SetEnabled sets updateCh in one goroutine, and
PeriodicDispatch.run accesses updateCh in another.
The race can be prevented by having SetEnabled pass updateCh to run.
Race detector output from `go test -race -run TestServer_RPC` in nomad/
```
==================
WARNING: DATA RACE
Write at 0x00c0001d3f48 by goroutine 75:
github.com/hashicorp/nomad/nomad.(*PeriodicDispatch).SetEnabled()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/periodic.go:468
+0x256
github.com/hashicorp/nomad/nomad.(*Server).revokeLeadership()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:724
+0x267
github.com/hashicorp/nomad/nomad.(*Server).leaderLoop.func1()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:131
+0x3c
github.com/hashicorp/nomad/nomad.(*Server).leaderLoop()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:163
+0x4dd
github.com/hashicorp/nomad/nomad.(*Server).monitorLeadership.func1()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:72
+0x6c
Previous read at 0x00c0001d3f48 by goroutine 515:
github.com/hashicorp/nomad/nomad.(*PeriodicDispatch).run()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/periodic.go:338
+0x177
Goroutine 75 (running) created at:
github.com/hashicorp/nomad/nomad.(*Server).monitorLeadership()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:70
+0x269
Goroutine 515 (running) created at:
github.com/hashicorp/nomad/nomad.(*PeriodicDispatch).SetEnabled()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/periodic.go:176
+0x1bc
github.com/hashicorp/nomad/nomad.(*Server).establishLeadership()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:231
+0x582
github.com/hashicorp/nomad/nomad.(*Server).leaderLoop()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:117
+0x82e
github.com/hashicorp/nomad/nomad.(*Server).monitorLeadership.func1()
/home/schmichael/go/src/github.com/hashicorp/nomad/nomad/leader.go:72
+0x6c
==================
```
IOPS have been modelled as a resource since Nomad 0.1 but has never
actually been detected and there is no plan in the short term to add
detection. This is because IOPS is a bit simplistic of a unit to define
the performance requirements from the underlying storage system. In its
current state it adds unnecessary confusion and can be removed without
impacting any users. This PR leaves IOPS defined at the jobspec parsing
level and in the api/ resources since these are the two public uses of
the field. These should be considered deprecated and only exist to allow
users to stop using them during the Nomad 0.9.x release. In the future,
there should be no expectation that the field will exist.
This PR fixes an edge case where we could GC an allocation that was in a
desired stop state but had not terminated yet. This can be hit if the
client hasn't shutdown the allocation yet or if the allocation is still
shutting down (long kill_timeout).
Fixes https://github.com/hashicorp/nomad/issues/4940
`currentExpiration` field is accessed in multiple goroutines: Stats and
renewal, so needs locking.
I don't anticipate high contention, so simple mutex suffices.
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.
Keep attempting to renew Vault token past locally recorded expiry, just
in case the token was renewed out of band, e.g. on another Nomad server,
until Vault returns an unrecoverable error.
Seems like the stats field is a micro-optimization that doesn't justify
the complexity it introduces. Removing it and computing the stats from
revoking field directly.
Vault's RenewSelf(...) API may return (nil, nil). We failed to check if
secret was nil before attempting to use it.
RenewSelf:
e3eee5b4fb/api/auth_token.go (L138-L155)
Calls ParseSecret:
e3eee5b4fb/api/secret.go (L309-L311)
If anyone has an idea on how to test this I didn't see any options. We
use a real Vault service, so there's no opportunity to mock the
response.
This adds constraints for asserting that a given attribute or value
exists, or does not exist. This acts as a companion to =, or !=
operators, e.g:
```hcl
constraint {
attribute = "${attrs.type}"
operator = "!="
value = "database"
}
constraint {
attribute = "${attrs.type}"
operator = "is_set"
}
```
This test expects 11 repeats of the same message emitted at intervals of
200ms; so we need more than 2 seconds to adjust for time sleep
variations and the like. So raising it to 3s here that should be
enough.
Fixes https://github.com/hashicorp/nomad/issues/4299
Upon investigating this case further, we determined the issue to be a race between applying `JobBatchDeregisterRequest` fsm operation and processing job-deregister evals.
Processing job-deregister evals should wait until the FSM log message finishes applying, by using the snapshot index. However, with `JobBatchDeregister`, any single individual job deregistering was applied accidentally incremented the snapshot index and resulted into processing job-deregister evals. When a Nomad server receives an eval for a job in the batch that is yet to be deleted, we accidentally re-run it depending on the state of allocation.
This change ensures that we delete deregister all of the jobs and inserts all evals in a single transactions, thus blocking processing related evals until deregistering complete.
The old logic for cancelling duplicate blocked evaluations by job id had
the issue where the newer evaluation could have additional node classes
that it is (in)eligible for that we would not capture. This could make
it such that cluster state could change such that the job would make
progress but no evaluation was unblocked.
Fix an issue in which the deployment watcher would fail the deployment
based on the earliest progress deadline of the deployment regardless of
if the task group has finished.
Further fix an issue where the blocked eval optimization would make it
so no evals were created to progress the deployment. To reproduce this
issue, prior to this commit, you can create a job with two task groups.
The first group has count 1 and resources such that it can not be
placed. The second group has count 3, max_parallel=1, and can be placed.
Run this first and then update the second group to do a deployment. It
will place the first of three, but never progress since there exists a
blocked eval. However, that doesn't capture the fact that there are two
groups being deployed.
This commit implements an allocation selection algorithm for finding
allocations to preempt. It currently special cases network resource asks
from others (cpu/memory/disk/iops).
* Migrated all of the old leader task tests and got them passing
* Refactor and consolidate task killing code in AR to always kill leader
tasks first
* Fixed lots of issues with state restoring
* Fixed deadlock in AR.Destroy if AR.Run had never been called
* Added a new in memory statedb for testing
