Commit Graph

129 Commits

Author SHA1 Message Date
7ac995cdde enhanced authBackend to support authReadTx 2022-04-07 05:35:13 +08:00
bfd5170f66 add a txPostLockHook into the backend
Previously the SetConsistentIndex() is called during the apply workflow,
but it's outside the db transaction. If a commit happens between SetConsistentIndex
and the following apply workflow, and etcd crashes for whatever reason right
after the commit, then etcd commits an incomplete transaction to db.
Eventually etcd runs into the data inconsistency issue.

In this commit, we move the SetConsistentIndex into a txPostLockHook, so
it will be executed inside the transaction lock.
2022-04-07 05:35:13 +08:00
836bd6bc3a fix WARNING: DATA RACE issue when multiple goroutines access the backend concurrently 2022-04-03 06:13:09 +08:00
edce939f6e add one more field storageVersion into StatusResponse
When performing the downgrade operation, users can confirm whether each member
is ready to be downgraded using the field 'storageVersion'. If it's equal to the
'target version' in the downgrade command, then it's ready to be downgraded;
otherwise, the etcd member is still in progress of processing the db file.
2022-03-18 07:04:44 +08:00
a0f26ff4ea server: Snapshot after cluster version downgrade 2022-02-21 15:48:00 +01:00
8c91d60a6f server: Switch to publishV3 2022-02-14 23:06:45 +01:00
0df768d2b1 server/etcdserver: fix oss-fuzz issue 42181 2022-02-14 10:59:41 +00:00
692b3c4cd7 server: Remove most of V2 API 2022-01-25 15:24:13 +01:00
ee5ef42c5c server: --enable-v2 and --enable-v2v3 is decomissioned 2022-01-14 13:19:30 +01:00
7d10899d7f server: Require either cluster version v3.6 or --experimental-enable-lease-checkpoint-persist to persist lease remainingTTL
To avoid inconsistant behavior during cluster upgrade we are feature
gating persistance behind cluster version. This should ensure that
all cluster members are upgraded to v3.6 before changing behavior.

To allow backporting this fix to v3.5 we are also introducing flag
--experimental-enable-lease-checkpoint-persist that will allow for
smooth upgrade in v3.5 clusters with this feature enabled.
2021-12-02 12:26:47 +01:00
e47c3c22d2 server: Move downgrade API logic into version package 2021-10-08 12:01:51 +02:00
1e5e57f268 server: Move downgrade detection code to version package 2021-10-08 10:41:37 +02:00
39f92a32ca server: Move member dir creation up and introduce Close method to bootstrap structs 2021-09-20 12:21:36 +02:00
4884e7d8cf server: Move wal bootstrap from cluster to storage 2021-09-20 12:21:35 +02:00
049e2d6ec0 server: Move raft up the bootstrap hierarchy 2021-09-20 12:20:19 +02:00
8b0d8ea2af server: Move cluster up the bootstrap hierarchy 2021-09-20 12:19:09 +02:00
7c8f7166e7 server: Move bootstraping backend from snapshot to bootstrapBackend 2021-09-20 12:17:33 +02:00
c97ab8f5e0 server: Move cluster up the bootstrap hierarchy 2021-09-20 12:07:41 +02:00
6a4ea70aef server: Move clusterID and nodeID up the bootstrap hierarchy 2021-09-20 12:06:18 +02:00
db06a4ab28 server: Move wal bootstrap up the hierarchy 2021-09-20 12:04:44 +02:00
aa0c050003 etcdserver: Add more hierarchy bootstap introducing a separate storage bootstrap step 2021-09-20 12:01:45 +02:00
66d05e5496 Try updating storage version immidietly after cluster version is set 2021-09-10 10:16:48 +02:00
ff3729c4d5 server: Implement storage schema migration to follow cluster version change and panic if unknown storage version is found
Storage version should follow cluster version. During upgrades this
should be immidiate as storage version can be always upgraded as storage
is backward compatible. During downgrades it will be delayed and will
require time for incompatible changes to be snapshotted.

As storage version change can happen long after cluster is running, we
need to add a step during bootstrap to validate if loaded data can be
understood by migrator.
2021-09-10 10:16:48 +02:00
9d81dde082 server: Extract notifier struct 2021-09-10 10:16:48 +02:00
87f1dc7e40 Fix a few typos 2021-09-03 16:09:09 -04:00
83a325ac46 server: Move all functions needed for storage bootstrap to storage package
This is prerequestite to move storage bootstrap, splitted to separate PR
to make it easier to review.
2021-08-03 13:09:15 +02:00
a0554a6bd3 etcdserver: Create AuthBackend interface 2021-07-20 18:09:53 +02:00
6cd3633543 etcdserver: Rename membershipStore to membershipBackend 2021-07-20 17:56:52 +02:00
2f31cc3fbc etcdserver: Create AlarmBackend interface 2021-07-20 17:53:44 +02:00
f6534f1e92 server: Move setting storage version to schema 2021-07-12 15:37:21 +02:00
5b6f4579fb server: Rename buckets to schema 2021-07-12 15:37:21 +02:00
5e40a8b00c server: Create storage package and move mvcc files to it 2021-07-12 15:37:21 +02:00
ef6c276fae etcdserver: Move server bootstraping to one file 2021-07-12 15:37:21 +02:00
9824cc96ed etcdserver: Fix typos in bootstrap 2021-07-08 13:37:19 +02:00
a72d4462fe etcdserver: Create boostrap wal functions 2021-07-07 23:36:35 +02:00
e75dfde4cb etcdserver: Move raft node start to just before newRaftNode 2021-07-07 23:26:37 +02:00
08935247a8 etcdserver: Create raftnode based on boostrapRaft struct 2021-07-07 17:38:56 +02:00
554777bba4 etcdserver: Extract boostrapRaft struct 2021-07-07 17:30:27 +02:00
7d39c3c655 etcdserver: Extract boostrapSnapshotter function 2021-07-07 16:14:33 +02:00
af0439490c etcdserver: Extract cluster boostrap functions 2021-07-07 16:04:52 +02:00
16b2a8b420 etcdserver: Prepare boostrap to split cluster setup 2021-07-07 15:56:43 +02:00
4dd9424d11 etcdserver: Extract boostrapBackend function 2021-07-07 15:46:38 +02:00
120cd5abe2 etcdserver: Extract etcdserver boostrap function 2021-07-07 15:40:54 +02:00
2db193fda1 etcdserver: Fix snapshot always nil 2021-07-07 13:29:45 +02:00
bf3e7033e9 etcdserver: Move Read/Update methods on Meta bucket to one place
There are still some left like compact keys, but they will require more
work to avoid circular dependency.
2021-07-05 13:23:53 +02:00
50507d5f3c etcdserver: Membership uses MembershipStorage interface instead of directly accessing Backend 2021-06-29 16:14:06 +02:00
e2740b4afa server,etcdutl: Preserve etcd version in backend allowing etcdutl to read it from snapshot 2021-06-25 14:06:56 +02:00
823f85dfc9 etcdserver: Move version monitor logic to separate module 2021-06-23 19:27:29 +02:00
184b0e5d49 add sleep interval 2021-05-24 16:22:00 +08:00
0a972a3f05 applyV2 should reapply on backend only once
During review of:  https://github.com/etcd-io/etcd/pull/12988 spotted
that PUT is actially writing to v3-backend.
If we are replaying WAL log, it might happened that backend's
applied_index is > than the WAL's log entry. In such situation we should
skip applying on backend V3.
I think both the methods (setVersion, setMembersAttributes) are in
practice idempotent so its not that 'serious' problem, but for
formal correctness adding the proper checks.
2021-05-18 20:55:05 +02:00