Execute retention lease syncs under system context #53838
Merged
Conversation
This file contains 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
The retention lease syncs need to occur under the system context, because they are internal actions executed on behalf of the user. Today we are relying on this happening for background syncs by virtue of the fact that the context the syncs are created under is the system context. This is due to these occurring on the cluster state applier thread. However, there are situations where this does not hold such as when a timed out cluster state publication occurs, and the node where the shard is allocated is the elected master node. In that case, the context will be empty due to the fact that we do not reschedule publication under the system context. Currently, doing so runs us into some troubles with losing the existing context, possibly dropping deprecation headers. We could copy that context over when marking the current context as the system context, but the implications of that require some more investigation. For now, we explicitly mark the retention lease syncs as executing under the system context, as this is situation that we can reason about.
jasontedor
added
>bug
:Distributed Indexing/Distributed
|
Pinging @elastic/es-distributed (:Distributed/Distributed) |
jasontedor
added a commit
that referenced
this pull request
Mar 20, 2020
The retention lease syncs need to occur under the system context, because they are internal actions executed on behalf of the user. Today we are relying on this happening for background syncs by virtue of the fact that the context the syncs are created under is the system context. This is due to these occurring on the cluster state applier thread. However, there are situations where this does not hold such as when a timed out cluster state publication occurs, and the node where the shard is allocated is the elected master node. In that case, the context will be empty due to the fact that we do not reschedule publication under the system context. Currently, doing so runs us into some troubles with losing the existing context, possibly dropping deprecation headers. We could copy that context over when marking the current context as the system context, but the implications of that require some more investigation. For now, we explicitly mark the retention lease syncs as executing under the system context, as this is situation that we can reason about.
jasontedor
added a commit
that referenced
this pull request
Mar 20, 2020
The retention lease syncs need to occur under the system context, because they are internal actions executed on behalf of the user. Today we are relying on this happening for background syncs by virtue of the fact that the context the syncs are created under is the system context. This is due to these occurring on the cluster state applier thread. However, there are situations where this does not hold such as when a timed out cluster state publication occurs, and the node where the shard is allocated is the elected master node. In that case, the context will be empty due to the fact that we do not reschedule publication under the system context. Currently, doing so runs us into some troubles with losing the existing context, possibly dropping deprecation headers. We could copy that context over when marking the current context as the system context, but the implications of that require some more investigation. For now, we explicitly mark the retention lease syncs as executing under the system context, as this is situation that we can reason about.
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.
The retention lease syncs need to occur under the system context, because they are internal actions executed on behalf of the user. Today we are relying on this happening for background syncs by virtue of the fact that the context the syncs are created under is the system context. This is due to these occurring on the cluster state applier thread. However, there are situations where this does not hold such as when a timed out cluster state publication occurs, and the node where the shard is allocated is the elected master node. In that case, the context will be empty due to the fact that we do not reschedule publication under the system context. Currently, doing so runs us into some troubles with losing the existing context, possibly dropping deprecation headers. We could copy that context over when marking the current context as the system context, but the implications of that require some more investigation. For now, we explicitly mark the retention lease syncs as executing under the system context, as this is situation that we can reason about.
Relates #48430
Relates #53751