Zhenyu Guo

#### Problem For RPC calls expecting response from server, we now issue a time-out task which triggers rpc response task with null response. Profiling reveals that this is one of...

server state consistency broken

4

i.e., in `server_state::check_consistency`, the last_drops contain the addresses that are assigned to primary and secondaries. This bug has been fixed once before the code merge, now it surfaces again.

this case is not handled yet.

#### Problem In our current design, L2 client can only be used to access one specific partition of an application. However, there are usage patterns that requires accessing all partitions...

### Problem As the core changes, it easily break the contract of the other language wrappers not in the main repo. This happens several times and it is really annoying....

log file ctor assertion failure

2

when run perf-test.sh for simple-kv under Linux, we got: 08:33:18.125 (1453105998125708440 3133) replica1.THREAD_POOL_REPLICATION_LONG7.0303000100123000: assertion expression: false 08:33:18.125 (1453105998125752230 3133) replica1.THREAD_POOL_REPLICATION_LONG7.0303000100123000: fail to get file size of /home/zhenyug/projects/rDSN/builder/bin/dsn.replication.simple_kv/data/replica1/1.0.simple_kv/plog/log.2.33556161.

replication log commit failure

3

``` C++ dassert(false, "cannot find handler for rpc code %d in %s", code, data_dir().c_str()); ``` in replication_app_base::dispatch_rpc_call. The root cause is that we store the rpc code as integers in...

Note: currently, we have to use different config files to do the test; after we enable per-app provider configuration, we can test them in one single process.

Note: currently, we have to use different config files to do the test; after we enable per-app provider configuration, we can test them in one single process.

different nodes may use different configurations (e.g., io loop mode), network providers. We need to change many core configurations to service node specific.