when BucketShardIncrementalSync encounters SYNCSTOP, it needs to call
try_update_high_marker() to advance past its marker. for sync modules
that don't support full sync, this allows them to resume incremental
sync from this position in the log when sync is reenabled
Signed-off-by: Casey Bodley <cbodley@redhat.com>
if (e.op == RGWModifyOp::CLS_RGW_OP_SYNCSTOP) {
ldout(sync_env->cct, 20) << "syncstop on " << e.timestamp << dendl;
syncstopped = true;
- entries_end = entries_iter; // dont sync past here
+ entries_end = std::next(entries_iter); // stop after this entry
break;
}
if (e.op == RGWModifyOp::CLS_RGW_OP_RESYNC) {