fix(gateway): keep ACME issuance off the startup path and bound ACME requests - #1262
Merged
Merged
Conversation
kvinwang
force-pushed
the
fix/gateway-startup-and-config
branch
from
September 24, 2026 08:42
8577bf8 to
960dcfc
Compare
kvinwang
changed the title
kvinwang
changed the title
…nd-config # Conflicts: # dstack/gateway/src/distributed_certbot.rs
kvinwang
added a commit
that referenced
this pull request
Sep 25, 2026
…in spelling PR #1262 removed certificate issuance from the startup path; the certificate-bootstrap cases now restart a node against a silent ACME directory with an uncertified domain configured and require it to listen. PR #1356 normalizes the domain in ForceReleaseCertLock and ListCertAttestations; tc-gw-certificat-002 releases a crashed node's lock with a non-normalized spelling and tc-gw-certificat-007 lists history under one. Signed-off-by: Kevin Wang <wy721@qq.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Two ways ACME work could block or hang the gateway.
ProxyInner::newcalledcertbot.init_all()beforeproxy::startbound any listener. It does ACME/DNS I/O per domain, andrequest_new_certslept a hard-coded 30s when another node held the cert lock. It is redundant: certificates already in the KV store are loaded into the resolver just above it, andstart_certbot_taskspawns a fullrenew_cert(None, false)pass that issues for domains without a certificate. Removedinit_all,init_domainandrequest_new_cert(single-caller each).ReqwestHttpClientuses reqwest's async client, which has no default timeout.ensure_acme_account/rotate_acme_credentialsrun under the cluster ACME lock with no outer bound, so a silent ACME directory held the lock until it expired. Each request now has a 30s timeout.Split out: peer HTTPS timeout (#1374), dead
external_portkey (#1375).Verification:
cargo test -p dstack-gateway -p certbotpasses; fmt and clippy clean.