Skip to content

[CORE-13370] archival: Fence spillover command #27714

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Sep 26, 2025
+37 −43
Merged

[CORE-13370] archival: Fence spillover command #27714

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
77 changes: 34 additions & 43 deletions src/v/cluster/archival/ntp_archiver_service.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -387,6 +387,16 @@ ntp_archiver::ntp_archiver(
}
}

archival_stm_fence ntp_archiver::emit_rw_fence() {
return {
.read_write_fence
= _parent.archival_meta_stm()->manifest().get_applied_offset(),
// Only use the rw-fence if the feature is enabled which requires
// major version upgrade.
.emit_rw_fence_cmd = emit_read_write_fence(_feature_table),
};
}

void ntp_archiver::log_collected_traces() noexcept {
try {
_rtclog.bypass_tracing([this] {
Expand Down Expand Up @@ -1053,13 +1063,7 @@ ss::future<> ntp_archiver::upload_until_term_change_legacy() {
std::optional<batch_result> result;
auto track_paused = _probe->register_archiver_on_hold(uploads_paused);
if (!uploads_paused) {
result = co_await upload_next_candidates(
archival_stm_fence{
.read_write_fence = fence,
// Only use the rw-fence if the feature is enabled which
// requires major version upgrade.
.emit_rw_fence_cmd = emit_read_write_fence(_feature_table),
});
result = co_await upload_next_candidates(emit_rw_fence());
}
if (result.has_value()) {
auto [compacted_upload_result, non_compacted_upload_result]
Expand Down Expand Up @@ -2687,13 +2691,7 @@ ss::future<> ntp_archiver::apply_archive_retention() {
co_return;
}

archival_stm_fence fence = {
.read_write_fence
= _parent.archival_meta_stm()->manifest().get_applied_offset(),
// Only use the rw-fence if the feature is enabled which requires
// major version upgrade.
.emit_rw_fence_cmd = emit_read_write_fence(_feature_table),
};
auto fence = emit_rw_fence();

std::optional<size_t> retention_bytes = ntp_conf.retention_bytes();
std::optional<std::chrono::milliseconds> retention_ms
Expand Down Expand Up @@ -2761,13 +2759,7 @@ ss::future<> ntp_archiver::garbage_collect_archive() {
if (!may_begin_uploads()) {
co_return;
}
archival_stm_fence fence = {
.read_write_fence
= _parent.archival_meta_stm()->manifest().get_applied_offset(),
// Only use the rw-fence if the feature is enabled which requires
// major version upgrade.
.emit_rw_fence_cmd = emit_read_write_fence(_feature_table),
};
auto fence = emit_rw_fence();
auto backlog = co_await _manifest_view->get_retention_backlog();
if (backlog.has_failure()) {
if (backlog.error() == cloud_storage::error_outcome::shutting_down) {
Expand Down Expand Up @@ -2996,7 +2988,7 @@ ss::future<> ntp_archiver::apply_spillover() {
if (!may_begin_uploads()) {
co_return;
}

archival_stm_fence fence = emit_rw_fence();
const auto manifest_upload_timeout = _conf->manifest_upload_timeout();
const auto manifest_upload_backoff = _conf->cloud_storage_initial_backoff();

Expand Down Expand Up @@ -3034,6 +3026,12 @@ ss::future<> ntp_archiver::apply_spillover() {
auto tail = [&]() {
cloud_storage::spillover_manifest tail(_ntp, _rev);
for (const auto& meta : manifest()) {
vlog(
_rtclog.trace,
"Adding segment {} to the spillover manifest that starts at "
"{}",
meta,
tail.get_start_offset().value_or(model::offset{}));
tail.add(meta);
// No performance impact since all writes here are
// sequential.
Expand Down Expand Up @@ -3090,6 +3088,13 @@ ss::future<> ntp_archiver::apply_spillover() {
auto deadline = ss::lowres_clock::now() + sync_timeout;

auto batch = _parent.archival_meta_stm()->batch_start(deadline, _as);
if (fence.emit_rw_fence_cmd) {
vlog(
_rtclog.debug,
"spillover, read-write fence: {}",
fence.read_write_fence);
batch.read_write_fence(fence.read_write_fence);
}
batch.spillover(spillover_meta);
if (manifest().get_archive_start_offset() == model::offset{}) {
vlog(
Expand All @@ -3111,12 +3116,15 @@ ss::future<> ntp_archiver::apply_spillover() {
_rtclog.warn,
"Failed to replicate spillover command: {}",
error.message());
break;
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

given the change in control flow here, i think a bit more exposition in the commit message would be helpful. i.e. the fact that we break out of the loop on replication failure, whether that's usually attributable to the fence, why it's ok to bail at this point, what happens next, etc.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done

} else {
vlog(
_rtclog.info,
"Uploaded spillover manifest: {}",
tail.get_manifest_path(remote_path_provider()));
}
// Reset fence for the next iteration
fence = emit_rw_fence();
}
}

Expand Down Expand Up @@ -3212,13 +3220,7 @@ ss::future<> ntp_archiver::apply_retention() {
if (!may_begin_uploads()) {
co_return;
}
archival_stm_fence fence = {
.read_write_fence
= _parent.archival_meta_stm()->manifest().get_applied_offset(),
// Only use the rw-fence if the feature is enabled which requires
// major version upgrade.
.emit_rw_fence_cmd = emit_read_write_fence(_feature_table),
};
auto fence = emit_rw_fence();
auto arch_so = manifest().get_archive_start_offset();
auto stm_so = manifest().get_start_offset();
if (arch_so != model::offset{} && arch_so != stm_so) {
Expand Down Expand Up @@ -3311,13 +3313,7 @@ ss::future<> ntp_archiver::garbage_collect() {
co_return;
}

archival_stm_fence fence = {
.read_write_fence
= _parent.archival_meta_stm()->manifest().get_applied_offset(),
// Only use the rw-fence if the feature is enabled which requires
// major version upgrade.
.emit_rw_fence_cmd = emit_read_write_fence(_feature_table),
};
archival_stm_fence fence = emit_rw_fence();

// If we are about to delete segments, we must ensure that the remote
// manifest is fully up to date, so that it is definitely not referring
Expand Down Expand Up @@ -3442,13 +3438,8 @@ ntp_archiver::find_reupload_candidate(

ssx::composite_abort_source cas{caller_as, _as};

archival_stm_fence rw_fence{
.read_write_fence
= _parent.archival_meta_stm()->manifest().get_applied_offset(),
// Only use the rw-fence if the feature is enabled which requires
// major version upgrade.
.emit_rw_fence_cmd = emit_read_write_fence(_feature_table),
};
archival_stm_fence rw_fence = emit_rw_fence();

if (!may_begin_uploads()) {
co_return find_reupload_candidate_result{};
}
Expand Down
3 changes: 3 additions & 0 deletions src/v/cluster/archival/ntp_archiver_service.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -466,6 +466,9 @@ class ntp_archiver {
static constexpr const char* segment_merger_ctx_label
= "adjacent_segment_merger";

/// Create a fence value for the next STM operation
archival_stm_fence emit_rw_fence();

/// Delete objects, return true on success and false otherwise
ss::future<bool>
batch_delete(std::vector<cloud_storage_clients::object_key> paths);
Expand Down