Repository navigation
Conversation
Both timers fire daily at midnight and can run during a Job. The cpuset partition keeps their CPU work off the benchmark cores, but their disk I/O and page cache churn reach the same disks that the guest drives write back to. Add both units to the maintenance drop-ins, so each waits for the runner's current Job under the job lock, and the runner takes no Job while the unit waits or runs.
Contributor
|
Bencher Report| Project | Bencher |
| Branch | u/ep/runner-quiet-maintenance |
| Testbed | intel-v1 |
Click to view all benchmark results
| Benchmark | Latency | Benchmark Result microseconds (µs) (Result Δ%) | Upper Boundary microseconds (µs) (Limit %) |
|---|---|---|---|
| Adapter::Json | 📈 view plot 🚷 view threshold | 5.14 µs(-0.81%)Baseline: 5.19 µs | 5.41 µs (95.14%) |
| Adapter::Magic (JSON) | 📈 view plot 🚷 view threshold | 4.97 µs(-1.08%)Baseline: 5.02 µs | 5.23 µs (94.95%) |
| Adapter::Magic (Rust) | 📈 view plot 🚷 view threshold | 26.25 µs(-2.17%)Baseline: 26.83 µs | 28.64 µs (91.63%) |
| Adapter::Rust | 📈 view plot 🚷 view threshold | 4.42 µs(-0.85%)Baseline: 4.46 µs | 4.59 µs (96.20%) |
| Adapter::RustBench | 📈 view plot 🚷 view threshold | 4.44 µs(-0.24%)Baseline: 4.45 µs | 4.58 µs (96.83%) |
They are not masked, as fstrim and the RAID units are not, so they stay in the list of unmasked units.
This branch has not been deployed
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
logrotate.timeranddpkg-db-backup.timerfire daily at midnight and can run during a Job. The cpuset partition keeps their CPU work off the benchmark cores, but their disk I/O and page cache churn reach the same disks that the guest drives write back to.This adds
logrotate.serviceanddpkg-db-backup.servicetoQUIET_UNITSinrunner_ops, socargo ops hostwrites the samezz-bencher-quiet.confdrop-in for each, andauditchecks that each unit's ownExecStart=is still the command the drop-in restates. Each unit then waits for the runner's current Job under the job lock, and the runner takes no Job while the unit waits or runs.Checked under systemd 255 on Ubuntu 24.04: with the job lock held, each unit stays in
activatingwith its marker in/run/bencher. The sandbox oflogrotate.service(ProtectSystem=full,PrivateDevices=true, and the others) does not get in the way: fd 3 is/dev/nulland fd 4 holds the lock. Once the lock is free, each unit runs to success and removes its marker.