VReplication: use new topo named locks and TTL override for workflow coordination

[mattlord]

Description

The VReplication and VDiff engines need to coordinate on workflows as the TableDiffer manipulates the associated workflow record (in the _vt.vreplication table) on its shard in order to initialize the diff (stopping the workflow, syncing up streams and snapshot positions, before restarting the workflow). In order to do so properly — meaning w/o impacting unrelated operations and thus able to use a lock that will not expire/be lost while work is in progress — they need a distributed lock in the topo server. Workflow information is, however, not stored in the topo server so there's no related record to lock. To address this gap, this PR adds support for named locks in the topo server. These are locks on an opaque name rather than a topo record/key. We then leverage this to lock the workflow using the unique name of targetkeyspace/workflowname.

Here's an example of their usage when running a vdiff with the local examples where the target keyspace is customer (which has 2 shards) and the workflow is commerce2customer:

❯ etcdctl watch --prefix=true /vitess/global/internal

PUT
/vitess/global/internal/named_locks/customer/commerce2customer/locks/7587879623616723911
{
  "Action": "vdiff",
  "HostName": "pslord.local",
  "UserName": "matt",
  "Time": "2024-06-25T14:51:21-04:00",
  "Status": "Running"
}
DELETE
/vitess/global/internal/named_locks/customer/commerce2customer/locks/7587879623616723911

PUT
/vitess/global/internal/named_locks/customer/commerce2customer/locks/7587879623616723958
{
  "Action": "vdiff",
  "HostName": "pslord.local",
  "UserName": "matt",
  "Time": "2024-06-25T14:51:21-04:00",
  "Status": "Running"
}
DELETE
/vitess/global/internal/named_locks/customer/commerce2customer/locks/7587879623616723958

There remains a general issue that the Keyspace locks taken during traffic switches — where we update various records in the topo related to routing rules, shard records, etc — can be lost after 30 seconds (by default, for etcd2topo). This PR also addresses this by adding a mechanism in the topo service, LockWithTTL, to override the default lock TTL for the topo implementation:

• For etcd it's 30 seconds configured via the --topo_etcd_lease_ttl flag
• For consul it's 15 seconds configured via the --topo_consul_lock_session_ttl flag
• For ZooKeeper there is no default as it does not support TTLs

We also now check to confirm that we are holding the locks between major operations.

Please see the RFC for additional details: #16269

Related Issue(s)

• Fixes: Bug Report: VDiff2 topo keyspace locks can be released too early #11811
• Implements: RFC: Improve Topology Server Locking #16269

Checklist

• "Backport to:" labels have been added if this change should be back-ported to release branches
• If this change is to be back-ported to previous releases, a justification is included in the PR description
• Tests were added or are not required
• Did the new or modified tests pass consistently locally and on CI?
• Documentation was added or is not required

[vitess-bot]

Review Checklist

Hello reviewers! 👋 Please follow this checklist when reviewing this Pull Request.

General

• Ensure that the Pull Request has a descriptive title.
• Ensure there is a link to an issue (except for internal cleanup and flaky test fixes), new features should have an RFC that documents use cases and test cases.

Tests

• Bug fixes should have at least one unit or end-to-end test, enhancement and new features should have a sufficient number of tests.

Documentation

• Apply the release notes (needs details) label if users need to know about this change.
• New features should be documented.
• There should be some code comments as to why things are implemented the way they are.
• There should be a comment at the top of each new or modified test to explain what the test does.

New flags

• Is this flag really necessary?
• Flag names must be clear and intuitive, use dashes (-), and have a clear help text.

If a workflow is added or modified:

• Each item in Jobs should be named in order to mark it as required.
• If the workflow needs to be marked as required, the maintainer team must be notified.

Backward compatibility

• Protobuf changes should be wire-compatible.
• Changes to _vt tables and RPCs need to be backward compatible.
• RPC changes should be compatible with vitess-operator
• If a flag is removed, then it should also be removed from vitess-operator and arewefastyet, if used there.
• vtctl command output order should be stable and awk-able.

[codecov]

Codecov Report

Attention: Patch coverage is 64.52991% with 83 lines in your changes missing coverage. Please review.

Project coverage is 68.69%. Comparing base (0db3577) to head (8de719a).

Files	Patch %	Lines
go/vt/vtctl/workflow/server.go	52.22%	43 Missing ⚠️
go/vt/topo/faketopo/faketopo.go	0.00%	8 Missing ⚠️
go/vt/topo/locks.go	77.14%	8 Missing ⚠️
go/vt/topo/consultopo/lock.go	53.84%	6 Missing ⚠️
go/vt/topo/etcd2topo/lock.go	40.00%	6 Missing ⚠️
go/vt/vttablet/tabletmanager/vdiff/table_differ.go	0.00%	5 Missing ⚠️
go/vt/topo/zk2topo/lock.go	0.00%	4 Missing ⚠️
...lient/command/vreplication/common/switchtraffic.go	0.00%	2 Missing ⚠️
go/vt/topo/memorytopo/lock.go	94.44%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #16260      +/-   ##
==========================================
- Coverage   68.71%   68.69%   -0.02%     
==========================================
  Files        1547     1548       +1     
  Lines      198287   198444     +157     
==========================================
+ Hits       136243   136330      +87     
- Misses      62044    62114      +70     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[GuptaManan100]

LGTM!

[deepthi]

Nice work!