Skip to content

fix(coprocessor): missing downgrade of scheduler error in tfhe-worker #1352

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
antoniupop merged 1 commit into from
Nov 18, 2025
Merged

fix(coprocessor): missing downgrade of scheduler error in tfhe-worker #1352

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
8 changes: 4 additions & 4 deletions coprocessor/fhevm-engine/scheduler/src/dfg/scheduler.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -483,7 +483,7 @@ async fn execute_partition(
for (h, i) in tx_inputs.iter_mut() {
if i.is_none() {
let Some(Ok(ct)) = res.get(h) else {
warn!(target: "scheduler", {transaction_id = ?tid },
warn!(target: "scheduler", {transaction_id = ?hex::encode(tid) },
"Missing input to compute transaction - skipping");
for nidx in dfg.graph.node_identifiers() {
let Some(node) = dfg.graph.node_weight_mut(nidx) else {
Expand All @@ -510,7 +510,7 @@ async fn execute_partition(
// Re-randomise inputs of the transaction - this also
// decompresses ciphertexts
if let Err(e) = re_randomise_transaction_inputs(tx_inputs, &tid, gpu_idx, cpk.clone()) {
error!(target: "scheduler", {transaction_id = ?tid, error = ?e },
error!(target: "scheduler", {transaction_id = ?hex::encode(tid), error = ?e },
"Error while re-randomising inputs");
for nidx in dfg.graph.node_identifiers() {
let Some(node) = dfg.graph.node_weight_mut(nidx) else {
Expand All @@ -536,7 +536,7 @@ async fn execute_partition(
// If re-randomisation is not available (e.g., on GPU),
// only decompress ciphertexts
if let Err(e) = decompress_transaction_inputs(tx_inputs, &tid, gpu_idx, cpk.clone()) {
error!(target: "scheduler", {transaction_id = ?tid, error = ?e },
error!(target: "scheduler", {transaction_id = ?hex::encode(tid), error = ?e },
"Error while decompressing inputs");
for nidx in dfg.graph.node_identifiers() {
let Some(node) = dfg.graph.node_weight_mut(nidx) else {
Expand Down Expand Up @@ -639,7 +639,7 @@ fn try_schedule_node(
cts.push(i);
} else {
// That should not be possible as we called the checker.
error!(target: "scheduler", { handle = ?node.result_handle }, "Computation missing inputs");
error!(target: "scheduler", { handle = ?hex::encode(&node.result_handle) }, "Computation missing inputs");
return;
}
}
Expand Down
5 changes: 5 additions & 0 deletions coprocessor/fhevm-engine/tfhe-worker/src/tfhe_worker.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -516,6 +516,11 @@ async fn upload_transaction_graph_results<'a>(
CoprocessorError::SchedulerError(SchedulerError::MissingInputs)
) {
uncomputable.push((result.handle.clone(), result.transaction_id.clone()));
// Make sure we don't mark this as an error since this simply means that the
// inputs weren't available when we tried scheduling these operations.
// Setting them as uncomputable will postpone them with an exponential backoff
// and they will be retried later.
continue;
}
}
WORKER_ERRORS_COUNTER.inc();
Expand Down
Loading