Revert "*: upgrade etcd to v3.4.3 (#1907)"

[nolouch]

This reverts commit 4c0c102.

What problem does this PR solve?

[2019/11/27 18:11:58.366 +08:00] [FATAL] [cluster.go:684] ["invalid downgrade; server version is lower than determined cluster version"] [current-server-version=3.3.0+git] [determined-cluster-version=3.4] [stack="go.etcd.io/etcd/etcdserver/api/membership.mustDetectDowngrade\n\t/Users/nolouch/csn/program/go/pkg/mod/go.etcd.io/etcd@v0.0.0-20190320044326-77d4b742cdbf/etcdserver/api/membership/cluster.go:684\ngo.etcd.io/etcd/etcdserver/api/membership.(*RaftCluster).SetVersion\n\t/Users/nolouch/csn/program/go/pkg/mod/go.etcd.io/etcd@v0.0.0-20190320044326-77d4b742cdbf/etcdserver/api/membership/cluster.go:497\ngo.etcd.io/etcd/etcdserver.(*applierV2store).Put\n\t/Users/nolouch/csn/program/go/pkg/mod/go.etcd.io/etcd@v0.0.0-20190320044326-77d4b742cdbf/etcdserver/apply_v2.go:96\ngo.etcd.io/etcd/etcdserver.(*EtcdServer).applyV2Request\n\t/Users/nolouch/csn/program/go/pkg/mod/go.etcd.io/etcd@v0.0.0-20190320044326-77d4b742cdbf/etcdserver/apply_v2.go:123\ngo.etcd.io/etcd/etcdserver.(*EtcdServer).applyEntryNormal\n\t/Users/nolouch/csn/program/go/pkg/mod/go.etcd.io/etcd@v0.0.0-20190320044326-77d4b742cdbf/etcdserver/server.go:1986\ngo.etcd.io/etcd/etcdserver.(*EtcdServer).apply\n\t/Users/nolouch/csn/program/go/pkg/mod/go.etcd.io/etcd@v0.0.0-20190320044326-77d4b742cdbf/etcdserver/server.go:1925\ngo.etcd.io/etcd/etcdserver.(*EtcdServer).applyEntries\n\t/Users/nolouch/csn/program/go/pkg/mod/go.etcd.io/etcd@v0.0.0-20190320044326-77d4b742cdbf/etcdserver/server.go:1349\ngo.etcd.io/etcd/etcdserver.(*EtcdServer).applyAll\n\t/Users/nolouch/csn/program/go/pkg/mod/go.etcd.io/etcd@v0.0.0-20190320044326-77d4b742cdbf/etcdserver/server.go:1075\ngo.etcd.io/etcd/etcdserver.(*EtcdServer).run.func8\n\t/Users/nolouch/csn/program/go/pkg/mod/go.etcd.io/etcd@v0.0.0-20190320044326-77d4b742cdbf/etcdserver/server.go:1020\ngo.etcd.io/etcd/pkg/schedule.(*fifo).run\n\t/Users/nolouch/csn/program/go/pkg/mod/go.etcd.io/etcd@v0.0.0-20190320044326-77d4b742cdbf/pkg/schedule/schedule.go:157"]

we cannot downgrade the master(etcd v3.4.3) to 3.0(etcd v.3.3.0) now, for the 3.2 branch, temporary revert the #1907
etcd downgrade: etcd-io/etcd#7308

What is changed and how it works?

revert

Check List

Tests

• Unit test

[codecov-io]

Codecov Report

Merging #1974 into master will increase coverage by 0.12%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #1974      +/-   ##
==========================================
+ Coverage   77.82%   77.94%   +0.12%     
==========================================
  Files         174      174              
  Lines       17414    17410       -4     
==========================================
+ Hits        13552    13571      +19     
+ Misses       2811     2790      -21     
+ Partials     1051     1049       -2

Impacted Files	Coverage Δ
pkg/metricutil/metricutil.go	90.62% <100%> (+4.51%)	⬆️
server/kv/etcd_kv.go	72.72% <0%> (-9.1%)	⬇️
server/api/util.go	63.46% <0%> (-3.85%)	⬇️
server/tso/tso.go	80.29% <0%> (-0.73%)	⬇️
client/client.go	70.02% <0%> (-0.42%)	⬇️
server/cluster.go	82.96% <0%> (-0.25%)	⬇️
server/schedule/operator_controller.go	84.6% <0%> (+0.18%)	⬆️
server/grpc_service.go	59.05% <0%> (+0.64%)	⬆️
pkg/btree/btree.go	87.65% <0%> (+0.8%)	⬆️
server/member/lease.go	93.75% <0%> (+1.56%)	⬆️
... and 7 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update a8b07a7...20fb1e4. Read the comment docs.

[overvenus]

What a pity!

[Luffbee]

LGTM

[nolouch]

/merge

[sre-bot]

/run-all-tests

[sre-bot]

@nolouch merge failed.

[nolouch]

Failed as expect.

[you06]

Does this fatal happens when a cluster's etcd data already upgraded to 3.4 but the cluster restart with 3.3? If so, the user who use master with #1907 branch will failed with this revert (some long term tests got failed for this). Will the etcd upgrade to 3.4 recently?