Refactor SnapshotReceiveHandler & Add UT #232
Conversation
| @@ -59,13 +58,14 @@ int HSHomeObject::SnapshotReceiveHandler::process_shard_snapshot_data(ResyncShar | |||
| shard_sb.chunk_id = chunk_id; | |||
|
|
|||
| homestore::MultiBlkId blk_id; | |||
| const auto hints = home_obj_.chunk_selector()->chunk_to_hints(chunk_id); | |||
| const auto hints = home_obj_.chunk_selector()->chunk_to_hints(shard_sb.chunk_id); | |||
| auto status = homestore::data_service().alloc_blks( | |||
| sisl::round_up(sizeof(shard_sb), homestore::data_service().get_blk_size()), hints, blk_id); | |||
| if (status != homestore::BlkAllocStatus::SUCCESS) { | |||
| LOGE("Failed to allocate blocks for shard {}", shard_meta.shard_id()); | |||
There was a problem hiding this comment.
not sure will we always baseline resync to a brand new member? if not, maybe we need to let emergency gc kick in here
There was a problem hiding this comment.
We prefer to clean up all the existing data in the PG at the start of a snapshot resync to reduce complexity. The cleanup logic will be implemented later.
koujl
force-pushed
the
baseline_resync
branch
from
766786f to
0847b98
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 @@
## baseline_resync #232 +/- ##
==================================================
Coverage ? 64.38%
==================================================
Files ? 33
Lines ? 2047
Branches ? 234
==================================================
Hits ? 1318
Misses ? 613
Partials ? 116 ☔ View full report in Codecov by Sentry. |
koujl
force-pushed
the
baseline_resync
branch
from
87f5849 to
8ebc276
Compare
SnapshotContextfromSnapshotReceiveHandlerto split constant and mutable membersSnapshotReceiveHandlerand fix some bugs