update config file using environment variable while docker create

Display table info

Fix memory leak, DoExecTask "arg" new but not delete.

context file path is wrong if pika is sharding mode

Add consensus last_log to command "info replication".

1. While follower processing leader hints, local searching by logic
index instead of binlog offset.

2. Check if enough follower while doing InternalCalCommittedIndex.

3. If slave node already exist, could still process HandleTrySyncRequest.

Replication-num defines how many followers in this single raft group

* Append dummyCmd when switch master
Co-authored-by: N刘军 <liujun19@ZBMAC-C02W12RWH.local>

pkcluster cmd add table option
del table with del slots

Remove conf server-id (no longer support)
Add limitation on conf consensus-level, only [0, 1, 2] is valid

Change keepalive time to be 2s, timeout time to be 20s as before

Basic Functional Test Done
Basic Unit Test Done

send trysync request will increase current term number by 1
leader will not commit log which is not equals to its current term
leader and follower will drop package if reveived term is smaller then current term
leader and folloer will update current term if received term is larger then current term

Stabilize context: applied_index

general case follower write binlog and apply binlog

Impl class SyncProgress to Manage slave node sync progress
Impl class MemLog to Manage memory log
Impl class StableLog to Manage Binlog files

Remove BinlogReaderManger
Separate class SlaveNode into single file

* feature: delete some Replication Manager API,reduce the procedure of find SyncMasterPartition shared_ptr and SyncSlavePartition  shared_ptr.

Move Binlog related op from PikaPartition into PikaStableLog
(SyncMasterPartition owns PikaStableLog)