[reconfigurator] Planner should wait for all MGS updates before proceeding to zones #9001
There are no files selected for viewing
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
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
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
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
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
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
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -12,6 +12,7 @@ use nexus_types::deployment::BlueprintHostPhase2DesiredContents; | |
| use nexus_types::deployment::PendingMgsUpdate; | ||
| use nexus_types::deployment::PendingMgsUpdateDetails; | ||
| use nexus_types::deployment::PendingMgsUpdateHostPhase1Details; | ||
| use nexus_types::deployment::planning_report::FailedMgsUpdateReason; | ||
| use nexus_types::inventory::BaseboardId; | ||
| use nexus_types::inventory::Collection; | ||
| use omicron_common::api::external::TufArtifactMeta; | ||
|
|
@@ -32,7 +33,7 @@ use tufaceous_artifact::ArtifactKind; | |
| /// | ||
| /// This is generated by the planning process whenever it also generates host | ||
| /// phase 1 updates. | ||
| #[derive(Debug, Clone, PartialEq, Eq)] | ||
| pub(crate) struct PendingHostPhase2Changes { | ||
| by_sled: BTreeMap<SledUuid, (M2Slot, BlueprintHostPhase2DesiredContents)>, | ||
| } | ||
|
|
@@ -270,24 +271,28 @@ pub(super) fn try_make_update( | |
| baseboard_id: &Arc<BaseboardId>, | ||
| inventory: &Collection, | ||
| current_artifacts: &TufRepoDescription, | ||
| ) -> Result< | ||
| Option<(PendingMgsUpdate, PendingHostPhase2Changes)>, | ||
| FailedMgsUpdateReason, | ||
| > { | ||
| let Some(sp_info) = inventory.sps.get(baseboard_id) else { | ||
| warn!( | ||
karencfv marked this conversation as resolved.
Outdated
Show resolved
Hide resolved
|
||
| log, | ||
| "cannot configure host OS update for board \ | ||
| (missing SP info from inventory)"; | ||
| baseboard_id, | ||
| ); | ||
| return Err(FailedMgsUpdateReason::SpNotInInventory); | ||
| }; | ||
|
|
||
| // Only configure host OS updates for sleds. | ||
| // | ||
| // We don't bother logging a return value of `None` for non-sleds, because | ||
| // we will never attempt to configure an update for them (nor should we). | ||
| // For the same reason, we do not return an error. | ||
| match sp_info.sp_type { | ||
| SpType::Sled => (), | ||
| SpType::Power | SpType::Switch => return Ok(None), | ||
| } | ||
|
|
||
| let Some(sled_agent) = inventory.sled_agents.iter().find(|sled_agent| { | ||
|
|
@@ -299,7 +304,7 @@ pub(super) fn try_make_update( | |
| (missing sled-agent info from inventory)"; | ||
| baseboard_id, | ||
| ); | ||
| return Err(FailedMgsUpdateReason::SledAgentInfoNotInInventory); | ||
| }; | ||
| let Some(last_reconciliation) = sled_agent.last_reconciliation.as_ref() | ||
| else { | ||
|
|
@@ -309,7 +314,7 @@ pub(super) fn try_make_update( | |
| (missing last reconciliation details from inventory)"; | ||
| baseboard_id, | ||
| ); | ||
| return Err(FailedMgsUpdateReason::LastReconciliationNotInInventory); | ||
| }; | ||
| let boot_disk = match &last_reconciliation.boot_partitions.boot_disk { | ||
| Ok(boot_disk) => *boot_disk, | ||
|
|
@@ -323,7 +328,9 @@ pub(super) fn try_make_update( | |
| baseboard_id, | ||
| "err" => err, | ||
| ); | ||
| return Err(FailedMgsUpdateReason::UnableToDetermineBootDisk( | ||
| err.to_string(), | ||
| )); | ||
| } | ||
| }; | ||
| let active_phase_2_hash = | ||
|
|
@@ -340,7 +347,11 @@ pub(super) fn try_make_update( | |
| "boot_disk" => ?boot_disk, | ||
| "err" => err, | ||
| ); | ||
| return Err( | ||
| FailedMgsUpdateReason::UnableToRetrieveBootDiskPhase2Image( | ||
| err.to_string(), | ||
| ), | ||
| ); | ||
| } | ||
| }; | ||
|
|
||
|
|
@@ -353,7 +364,7 @@ pub(super) fn try_make_update( | |
| (inventory missing current active host phase 1 slot)"; | ||
| baseboard_id, | ||
| ); | ||
| return Err(FailedMgsUpdateReason::ActiveHostPhase1SlotNotInInventory); | ||
| }; | ||
|
|
||
| // TODO-correctness What should we do if the active phase 1 slot doesn't | ||
|
|
@@ -376,7 +387,9 @@ pub(super) fn try_make_update( | |
| "active_phase_1_slot" => ?active_phase_1_slot, | ||
| "boot_disk" => ?boot_disk, | ||
| ); | ||
| return Err( | ||
| FailedMgsUpdateReason::ActiveHostPhase1SlotBootDiskMismatch, | ||
| ); | ||
| } | ||
|
|
||
| let Some(active_phase_1_hash) = inventory | ||
|
|
@@ -390,7 +403,7 @@ pub(super) fn try_make_update( | |
| baseboard_id, | ||
| "slot" => ?active_phase_1_slot, | ||
| ); | ||
| return Err(FailedMgsUpdateReason::ActiveHostPhase1HashNotInInventory); | ||
| }; | ||
|
|
||
| let Some(inactive_phase_1_hash) = inventory | ||
|
|
@@ -407,7 +420,9 @@ pub(super) fn try_make_update( | |
| baseboard_id, | ||
| "slot" => ?active_phase_1_slot.toggled(), | ||
| ); | ||
| return Err( | ||
| FailedMgsUpdateReason::InactiveHostPhase1HashNotInInventory, | ||
| ); | ||
| }; | ||
|
|
||
| let mut phase_1_artifacts = Vec::with_capacity(1); | ||
|
|
@@ -433,7 +448,7 @@ pub(super) fn try_make_update( | |
| (no phase 1 artifact)"; | ||
| baseboard_id, | ||
| ); | ||
| return Err(FailedMgsUpdateReason::NoMatchingArtifactFound); | ||
| } | ||
| (_, []) => { | ||
| warn!( | ||
|
|
@@ -442,7 +457,7 @@ pub(super) fn try_make_update( | |
| (no phase 2 artifact)"; | ||
| baseboard_id, | ||
| ); | ||
| return Err(FailedMgsUpdateReason::NoMatchingArtifactFound); | ||
|
There was a problem hiding this comment. Could we add some context to Another thought, although this might be nonsense after reading the rest of the PR - we could potentially have different error types for each kind of update, so we could have more host-specific error variants, and then combine the error types in a higher-level enum? There was a problem hiding this comment. That makes sense, I'll address this in the follow up PR that will break up |
||
| } | ||
| // "TUF is broken" cases: have multiple of one or the other. This | ||
| // should be impossible unless we shipped a TUF repo with multiple | ||
|
|
@@ -457,7 +472,7 @@ pub(super) fn try_make_update( | |
| "num-phase-1-images" => phase_1_artifacts.len(), | ||
| "num-phase-2-images" => phase_2_artifacts.len(), | ||
| ); | ||
| return Err(FailedMgsUpdateReason::TooManyMatchingArtifacts); | ||
| } | ||
| }; | ||
|
|
||
|
|
@@ -469,7 +484,7 @@ pub(super) fn try_make_update( | |
| // this sled will fail to boot if it were rebooted now.) | ||
| if active_phase_2_hash == phase_2_artifact.hash { | ||
| debug!(log, "no host OS update needed for board"; baseboard_id); | ||
| return Ok(None); | ||
| } | ||
|
|
||
| // Before we can proceed with the phase 1 update, we need sled-agent to | ||
|
|
@@ -485,7 +500,7 @@ pub(super) fn try_make_update( | |
| phase_2_artifact, | ||
| ); | ||
|
|
||
| Ok(Some(( | ||
| PendingMgsUpdate { | ||
| baseboard_id: baseboard_id.clone(), | ||
| sp_type: sp_info.sp_type, | ||
|
|
@@ -505,7 +520,7 @@ pub(super) fn try_make_update( | |
| artifact_version: phase_1_artifact.id.version.clone(), | ||
| }, | ||
| pending_host_phase_2_changes, | ||
| ))) | ||
| } | ||
|
|
||
| #[cfg(test)] | ||
|
|
||
Oops, something went wrong.
Oops, something went wrong.
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.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We don't necessarily need to change this on this PR, but this type is pretty beefy. I wonder if it would be clearer for us to define an enum with three states; something like (names are hard)
so details like "
Ok(None)means there's no update needed" are more explicit?There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I like that idea! I've left some TODOs to finish up in a follow up PR in cb197db