Merge remote-tracking branch 'puppet/7.x' into doc-7.y #3865
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
This updates lein-ezbake to 2.4.2 to add build target support for ubuntu 2204
…dd-ubuntu-2204 (PDB-5636) update ezbake to 2.4.2
This adds ubuntu 2204/jammy to the set of supported puppetdb versions starting now.
…uthentication-configuration-a-go-go (PDB-5636) add ubuntu 2204 as supported platform
I think this carried over from the last refactor; since user is null here, we end up printing "Disconnecting all null connections" in the log. Changed it to non-migrator, since I believe that is effectively what we are doing here.
Previous behavior would cause compilers to be disconnected from postgresql ever time the primary's pe-puppetdb was restarted. This was because the restart invokes the migration code path, which, in PE with a migrator-account set leads to connection blocking even when the migration itself is ultimately a noop. Now it tests whether any migrations are pending before calling call-with-connections-blocked-during-migration. This change also means that update-schema will return nil in this case, and initialize-schema will return false, which is a change from previous noop migrations where initialize-schema would have returned true.
…-connections-when-no-migrations-pending (PE-36120) Do not kill connections when no migrations pending
Change PQL expr-or and expr-and grammar rules to use a recursive
definition.
By using instaparse's `{ }` operator (meaning zero or more) in
self-referential rules the grammar is ambiguous.
a = 1 or b = 2 or c = 3 can be parsed as something that looks like
[:or
a = 1
b = 2
c = 3]
Or it could be parsed as
[:or
a = 1
[:or
b = 2
c = 3]]
Both are valid given the self-referential grammar rule, but as clauses
compound the number of valid parsings grows exponentially significantly
affecting performance. As few as 16 or clauses can cause an OOM with 1GB
heap.
The recursive grammar definition means only one (the second example
above) version is valid. Every and/or operator will have two children.
The patch to the pql-grammar.ebnf took care of the ambiguity causing exponential parsing of deep OR clauses, but produces a nested AST form that differs from the flat [or 1 2 3 4] structure that was previously being handed off to the query engine. Nick Lewis had suggested an additional patch to transform.clj which would restructure the AST back to the form that the PQL parser was previously producing so that the rest of the system isn't blind-sided by a differing query structure. Since the issue we were fixing was in the initial parse of the PQL, taken together, these patches should fix the slow parse time without introducing any other novel behavior in the query engine.
This was removed from trapperkeeper due to a CVE, but it is used in our testing. (Originally this was in main, but the same adjustment to tk removing clj-yaml as a dependency made it's way into clj-parent 5.x and we need this in puppetdb 7.x now as well.)
(PDB-5643) Use recursive PQL rules for and/or
…zbake-to-2.5.0-for-logback (PE-36184) Update ezbake to 2.5.0 in prep for logback 1.3.x bump
This updates ezbake to 2.5.2, which does two things: * it enables builds of the el9 platform * it adds in tzdata-java as a dependency on the el platforms tzdata-java was strangely removed as a dependency from openjdk, so this adds it in as a dependendency to ensure we don't release a broken package.
…te-ezbake (maint) update ezbake to 2.5.2
They've been deprecated, "set --local path/without" instead.
This avoids conflating depdendencies, so that we should use exactly the set for each server set that'll be in running when deployed.
This started breaking checkouts of say 6.x.
…ppetserver-jvm-for-int-tests (PDB-5572) Run separate puppetserver JVM for :integration tests
…zbake-for-rhel9 (PDB-5670, PDB-5668) Update ezbake and beaker-hostgenerator for rhel9 support
contains a fix to allow package builds on el9
Don't fail to log just because the producer timestamp is nil.
Don't assume it won't be nil; sometimes it is.
…command-error-handling (PDB-5633) Handle nil producer timestamp and sql states during errors
…-el-9 (PDB-5667, PDB-5671) add el-9 to supported platforms
…se-notes (PDB-5675) Add 7.14.0 release notes
* puppet/7.x: (maint) Update version to 7.14.1-SNAPSHOT (maint) Update version to 7.14.0 for release (PDB-5675) Add 7.14.0 release notes (PDB-5667, PDB-5671) add el-9 to supported platforms (PDB-5633) exec-command: handle nil getSQLState values (PDB-5633) log-command-processed-message: handle nil producer-ts (maint) Update ezbake to 2.5.3 (maint) exclude beaker gems from integration/rspec tests (PDB-5670, PDB-5668) Update ezbake and beaker-hostgenerator for rhel9 support Set clj-parent=5.5.0 (PDB-5572) Don't limit :integration test puppetserver clone depth (PDB-5572) Run puppetserver in its own JVM in :integration tests (PDB-5572) config-puppet-test-ref: avoid bundle --path and --without Set clj-parent=5.4.0 (maint) update ezbake to 2.5.2 (PE-36184) Update ezbake to 2.5.0 in prep for logback 1.3.x bump (maint) Add clj-yaml as a dev dependency (PDB-5643) Transform back to flattened OR AST expressions (PDB-5643) Add a regression test for the parse (PDB-5643) Use recursive PQL rules for and/or Set clj-parent=5.3.9 Set clj-parent=5.3.8 (i18n) Update strings in puppetdb.pot file (PE-36120) Do not kill connections when no migrations are pending (maint) Don't print null user in disconnecting message (PDB-5636) add ubuntu 2204 as supported platform (PDB-5636) update ezbake to 2.4.2
|
|
jonathannewman
approved these changes
Aug 24, 2023
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.
No description provided.