[internal-dns] Cleanup server/storage on successful test run #993
Conversation
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
davepacheco
approved these changes
Apr 28, 2022
Collaborator
davepacheco
left a comment
davepacheco
left a comment
There was a problem hiding this comment.
Nice! Thanks for doing this.
iximeow
added a commit
that referenced
this pull request
Jan 8, 2026
Propolis changes are pretty important, it's unfortunate I'd overlooked driving the changes through to Omicron. The full list of changes: * Do not lose wakeups for block device workers (#973) * fix arg for dtrace script (#978) - no changes to Propolis * Add support for programmable SMBIOS Type 1 table (#977) - unused in Nexus * NVMe reset can discard request Permits early (#983) * Wire up viona notify fast path (#754) * distinguish probes across NVMe devices (#993) * update dropshot to 0.16.6, dropshot-api-manager to 0.3.0 (#992) * distinguish file backend thread names across backend instances (#997) * bump softnpu (#1000) * block/file: do pread/pwrite from Propolis heap instead of VM memory (#985) * bins: intial CPU binding support (#991) At least some of these are fixes for bugs we've seen in the last month or two (#983 made for a very coreful dogfood update as Propolises were stopped), many are either unused by Nexus (#977, #987, #993, #997) or just not reachable from a Nexus-configured VM *yet* (#973, #985). Both #991 and #754 are effectful today. The initial CPU binding in Propolis is such that if a VM would use more than half of a sled's CPUs, we explicitly bind vCPU threads to uppermost CPUs 1:1. This helps keep VM exit/reentry quick, mostly interesting while under high I/O load. In the limit this binding should be chosen by Nexus, and probably should be applied for all VMs. Additionally, #754 affects all VM NICs. It's just to avoid going out to Propolis and back for what really can be handled in-kernel; there should be no behavioral change here. Finally, while there *is* a new `propolis-server` API version, and I have sled-agent using it, I'm intentionally not plumbing that through to Nexus. There isn't anything valuable for Nexus to do with it, so it would be a new sled-agent API rev to include the new Propolis API type .. for no reason. Instead, I've opted to have sled-agent do the lossless conversion from older Propolis spec as provided by Nexus, to newer Propolis spec. We'll definitely have more changes to the Propolis API that will go through to Nexus, such as aforementioned CPU binding assignments, so I figure this will get more normalized sooner than later.
askfongjojo
pushed a commit
that referenced
this pull request
Jan 9, 2026
Propolis changes are pretty important, it's unfortunate I'd overlooked driving the changes through to Omicron. The full list of changes: * Do not lose wakeups for block device workers (#973) * fix arg for dtrace script (#978) - no changes to Propolis * Add support for programmable SMBIOS Type 1 table (#977) - unused in Nexus * NVMe reset can discard request Permits early (#983) * Wire up viona notify fast path (#754) * distinguish probes across NVMe devices (#993) * update dropshot to 0.16.6, dropshot-api-manager to 0.3.0 (#992) * distinguish file backend thread names across backend instances (#997) * bump softnpu (#1000) * block/file: do pread/pwrite from Propolis heap instead of VM memory (#985) * bins: intial CPU binding support (#991) At least some of these are fixes for bugs we've seen in the last month or two (#983 made for a very coreful dogfood update as Propolises were stopped), many are either unused by Nexus (#977, #987, #993, #997) or just not reachable from a Nexus-configured VM *yet* (#973, #985). Both #991 and #754 are effectful today. The initial CPU binding in Propolis is such that if a VM would use more than half of a sled's CPUs, we explicitly bind vCPU threads to uppermost CPUs 1:1. This helps keep VM exit/reentry quick, mostly interesting while under high I/O load. In the limit this binding should be chosen by Nexus, and probably should be applied for all VMs. Additionally, #754 affects all VM NICs. It's just to avoid going out to Propolis and back for what really can be handled in-kernel; there should be no behavioral change here. Finally, while there *is* a new `propolis-server` API version, and I have sled-agent using it, I'm intentionally not plumbing that through to Nexus. There isn't anything valuable for Nexus to do with it, so it would be a new sled-agent API rev to include the new Propolis API type .. for no reason. Instead, I've opted to have sled-agent do the lossless conversion from older Propolis spec as provided by Nexus, to newer Propolis spec. We'll definitely have more changes to the Propolis API that will go through to Nexus, such as aforementioned CPU binding assignments, so I figure this will get more normalized sooner than later.
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.
Fixes #990
Easy repro:
Before: On success, the temporary directory grew. After: It's empty.