fix(dup): don't GC by valid_start_offset during duplication & add app_name to replica_base

[neverchanje]

What problem this PR solved

• In some situations, we need metrics of a certain table on a certain pegasus node. Like that table-level get-latency does. The metric name usually uses "app_name" rather than "gpid" for readability to indicate the table it belongs to. So in this PR, I added "app_name" to replica_base. Therefore classes like 'pegasus_server_impl' can have knowledge of its corresponding app_name.

struct replica_base
{
-   replica_base(gpid id, string_view name) : _gpid(id), _name(name) {}
+   replica_base(gpid id, string_view name, string_view app_name)
+      : _gpid(id), _name(name), _app_name(app_name)
+   {
+   }

• There is also a data durability issue in plog-GC during duplication. Because of historical issues, we use "app_init_offset" to trigger GC for the logs with file offset before "app_init_offset". Some not-confirmed mutations will be deleted due to this reason. So here we set valid_start_offset to 0 to avoid unexpected GC.

        int64_t valid_start_offset = _app->init_info().init_offset_in_private_log;
        if (min_confirmed_decree >= 0) {
+           // Do not rely on valid_start_offset for GC during duplication.
+           // cleanable_decree is the only GC trigger.
+           valid_start_offset = 0;
            if (min_confirmed_decree < last_durable_decree) {
                cleanable_decree = min_confirmed_decree;
            }
        } else if (is_duplicating()) {
            // unsure if the logs can be dropped, because min_confirmed_decree
            // is currently unavailable
            return;
        }
        plog->garbage_collection(get_gpid(), cleanable_decree, valid_start_offset);

Manual Test:

1. Start two onebox clusters, A and B, then set up duplication A->B.
2. Loading writes to A, ensure A is duplicating to B.
3. Shut down B's replica-server. Ensure A is failing to duplicate (pending_mutations_count increases)
   
4. Ensure no GC of the private log after a few minutes.

   grep gc_private onebox/replica* | grep removed

[foreverneverer]

add app_name to replica_base let the pegasus some code compile error(pegasus and rdsn all latest branch), such as pegasus_mutation_duplicator_test.cpp, it call old version replica_base constructor:

void test_duplicate()
    {
        replica_base replica(dsn::gpid(1, 1), "fake_replica");
        auto duplicator = new_mutation_duplicator(&replica, "onebox2", "temp");
        duplicator->set_task_environment(&_env);
....
    }

https://github.com/XiaoMi/pegasus/blob/fad0c4f44534ca8daa991b10fee998819974b8d1/src/server/test/pegasus_mutation_duplicator_test.cpp#L33-L36