rework recovery #152
rework recovery #152
Conversation
JacksonYao287
force-pushed
the
rework-recovery
branch
from
f34e952 to
8d7d0ab
Compare
Codecov ReportAttention: Patch coverage is
❗ Your organization needs to install the Codecov GitHub app to enable full functionality. Additional details and impacted files@@ Coverage Diff @@
## main #152 +/- ##
==========================================
- Coverage 75.56% 75.36% -0.21%
==========================================
Files 30 30
Lines 1326 1319 -7
Branches 139 137 -2
==========================================
- Hits 1002 994 -8
- Misses 244 245 +1
Partials 80 80 ☔ View full report in Codecov by Sentry. |
JacksonYao287
force-pushed
the
rework-recovery
branch
from
ff9621f to
45e963d
Compare
| @@ -162,6 +162,7 @@ TEST_F(ShardManagerTestingRecovery, ShardManagerRecovery) { | |||
| _home_object.reset(); | |||
| LOGI("restart home_object"); | |||
| _home_object = homeobject::init_homeobject(std::weak_ptr< homeobject::HomeObjectApplication >(app)); | |||
| std::this_thread::sleep_for(std::chrono::seconds{5}); | |||
There was a problem hiding this comment.
could we use pg->repl_dev_->get_leader_id() to check in L190 and sleep if not the leader?
There was a problem hiding this comment.
I use 5s because the election_time_out is 3.2s. https://github.com/eBay/nuraft_mesg/blob/7e8cdeb261270642dac186902351da2174d7f077/src/lib/manager_impl.cpp#L25C1-L25C72
i want to refactor this two test case here , since they are much similiar and have a lot of duplicated logic as
can we do this in a separate refactor PR?
recovery pg and shard after homestore is fully restarted
this io_uring problem is tracked in this issue