Crdb 29662 #2
Closed
Crdb 29662 #2
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
Release note: None
Informs cockroachdb#128308 Epic: CRDB-37515 Release note: None
Changes: * [`5627dafd`](cockroachdb/pebble@5627dafd) db: fix compaction bounds expansion to avoid empty key comparisons * [`9a4f0c8c`](cockroachdb/pebble@9a4f0c8c) sstable: fix value block closure allocation * [`791b3749`](cockroachdb/pebble@791b3749) colblk: add PrefixBytesIter for stateful key synthesis * [`3419a647`](cockroachdb/pebble@3419a647) colblk: add UnsafeIntegerSlice microbenchmark * [`f1d8c9a0`](cockroachdb/pebble@f1d8c9a0) db: export MakeTrailer * [`ba34632e`](cockroachdb/pebble@ba34632e) colblk: add PrefixBytes.SetNextAt * [`64a28005`](cockroachdb/pebble@64a28005) colblk: inline SharedPrefix in PrefixBytes.AppendAt * [`aacee0a0`](cockroachdb/pebble@aacee0a0) colblk: use pointer receiver for RawBytes, PrefixBytes * [`daeeafe0`](cockroachdb/pebble@daeeafe0) colblk: add PrefixBytes.AppendAt * [`65445b8a`](cockroachdb/pebble@65445b8a) rowblk: add cockroach microbenchmarks * [`cda4471a`](cockroachdb/pebble@cda4471a) colblk: fix UintBuilder's construction of constant column with default * [`2e4bf3a6`](cockroachdb/pebble@2e4bf3a6) devtools: use separate module for dev tool dependencies * [`62cd2a25`](cockroachdb/pebble@62cd2a25) sstable: extract formatWriterMetadata test helper Release note: none. Epic: none.
128754: kvserverpb: add RaftMessageRequest.UsingRAC2Protocol r=kvoli a=sumeerbhola Informs cockroachdb#128309 Epic: CRDB-37515 Release note: None Co-authored-by: sumeerbhola <[email protected]>
The CRDB data mount is different on azure when compared to AWS and GCE. We use `sda1` instead of `data1` in this case. Fixes: cockroachdb#128364. Release note: None
Release note (enterprise change): Update certain error messages to refer to the "stable" docs tree rather than an explicit version. Fixes: cockroachdb#128611
128839: go.mod: bump Pebble to 5627dafd7951 r=jbowens a=itsbilal Changes: * [`5627dafd`](cockroachdb/pebble@5627dafd) db: fix compaction bounds expansion to avoid empty key comparisons * [`9a4f0c8c`](cockroachdb/pebble@9a4f0c8c) sstable: fix value block closure allocation * [`791b3749`](cockroachdb/pebble@791b3749) colblk: add PrefixBytesIter for stateful key synthesis * [`3419a647`](cockroachdb/pebble@3419a647) colblk: add UnsafeIntegerSlice microbenchmark * [`f1d8c9a0`](cockroachdb/pebble@f1d8c9a0) db: export MakeTrailer * [`ba34632e`](cockroachdb/pebble@ba34632e) colblk: add PrefixBytes.SetNextAt * [`64a28005`](cockroachdb/pebble@64a28005) colblk: inline SharedPrefix in PrefixBytes.AppendAt * [`aacee0a0`](cockroachdb/pebble@aacee0a0) colblk: use pointer receiver for RawBytes, PrefixBytes * [`daeeafe0`](cockroachdb/pebble@daeeafe0) colblk: add PrefixBytes.AppendAt * [`65445b8a`](cockroachdb/pebble@65445b8a) rowblk: add cockroach microbenchmarks * [`cda4471a`](cockroachdb/pebble@cda4471a) colblk: fix UintBuilder's construction of constant column with default * [`2e4bf3a6`](cockroachdb/pebble@2e4bf3a6) devtools: use separate module for dev tool dependencies * [`62cd2a25`](cockroachdb/pebble@62cd2a25) sstable: extract formatWriterMetadata test helper Release note: none. Epic: none. Co-authored-by: Bilal Akhtar <[email protected]>
There was a bug where comments did not show up in the output of the show create all types command. The issue was that we did not have any code to get the comment and add it to the create type descriptor. I added these changes and also did a bit of cleanup/removed duplicate code in the process. Fixes: cockroachdb#126005 Release note (bug fix): SHOW CREATE ALL TYPES now shows corresponding type comments in it's output.
It performed a very basic upgrade test and no longer serves a purpose. Epic: none Release note: None
128842: backup/restore: change documentation links to "stable" r=benbardin a=benbardin Release note (enterprise change): Update certain error messages to refer to the "stable" docs tree rather than an explicit version. Fixes: cockroachdb#128611 Co-authored-by: Ben Bardin <[email protected]>
- [x] Add version key constant for new release (e.g. `V24.2`), equal to `Latest` - [x] Update `PreviousRelease` constant - [x] Add start version (e.g. `V24.2Start` with version `24.1-2`) - [x] Update `pkg/build/version.txt` to the new version (e.g. `v24.2.0-alpha.00000000`) - [x] Add mixed version logictest config for the replaced version - [x] Update the `scplan` rules in `pkg/sql/schemachanger/scplan/internal/rules` - [x] Create new roachtest fixtures for the previous version - [x] Create new SQL bootstrap data - [x] Update releases file This change does not update the version skipping logic; in a follow-up change we will enable direct 24.1->24.3 upgrade without a special env flag. Epic: REL-1163 Release note: None
128616: clusterversion: create 24.3 version r=RaduBerinde a=RaduBerinde - [x] Add version key constant for new release (e.g. `V24.2`), equal to `Latest` - [x] Update `PreviousRelease` constant - [x] Add start version (e.g. `V24.2Start` with version `24.1-2`) - [x] Update `pkg/build/version.txt` to the new version (e.g. `v24.2.0-alpha.00000000`) - [x] Add mixed version logictest config for the replaced version - [x] Update the `scplan` rules in `pkg/sql/schemachanger/scplan/internal/rules` - [x] Create new roachtest fixtures for the previous version - [x] Create new SQL bootstrap data - [x] Update releases file Epic: REL-1163 Release note: None Fixes cockroachdb#128747 128706: rac2: specify lock ordering, and add ReplicaSet r=kvoli a=sumeerbhola Informs cockroachdb#128308 Epic: CRDB-37515 Release note: None Co-authored-by: Radu Berinde <[email protected]> Co-authored-by: sumeerbhola <[email protected]>
These check constraints are added temporarily while adding a non-nullable column. We have seen flakes caused by the workload trying to refer to them. Release note: None
The update_heavy test had failures because the ycsb workload was using column families, which aren't supported for LDR. Release note: none Fixes: cockroachdb#128740
…actions This will avoid errors of the form: "cannot perform other schema changes in the same transaction as a primary key change." Release note: None
fallBackIfDescColInRowLevelTTLTables is no longer needed as of e5b30e4. Release note: None
…achdb#128855 128833: workload/schemachange: avoid commmenting on transient check constraints r=rafiss a=rafiss These check constraints are added temporarily while adding a non-nullable column. We have seen flakes caused by the workload trying to refer to them. informs cockroachdb#128615 Release note: None 128834: workload/schemachange: only allow PK change in single statement transactions r=rafiss a=rafiss This will avoid errors of the form: "cannot perform other schema changes in the same transaction as a primary key change." informs: cockroachdb#128615 Release note: None 128851: roachtest: Fix ycsb workload to avoid column families r=dt a=navsetlur The update_heavy test had failures because the ycsb workload was using column families, which aren't supported for LDR. Release note: none Fixes: cockroachdb#128740 128855: schemachanger: remove unneeded function r=rafiss a=rafiss fallBackIfDescColInRowLevelTTLTables is no longer needed as of e5b30e4. Epic: None Release note: None Co-authored-by: Rafi Shamim <[email protected]> Co-authored-by: Naveen Setlur <[email protected]>
This patch attempts to deflake the changefeed distribution strategy unit tests by replacing all the manually constructed `ALTER TABLE ... RELOCATE` statements with a single statement that uses a select query. Release note: None
running workload with fresh env avoids killing it during stop/put/start cycles. Release note: none. Epic: none.
128084: sql: SHOW CREATE ALL TYPES doesn't work with comment on r=Dedej-Bergin a=Dedej-Bergin There was a bug where comments did not show up in the output of the show create all types command. The issue was that we did not have any code to get the comment and add it to the create type descriptor. I added these changes and also did a bit of cleanup/removed duplicate code in the process. Fixes: cockroachdb#126005 Release note (bug fix): SHOW CREATE ALL TYPES now shows corresponding type comments in it's output. 128715: telemetry: remove expectation of no force-custom counter r=mgartner,kyle-a-wong a=dhartunian This test flakes in cases where we run a query and expect the `sql.plan.type.force-custom` to not get incremented. This can't be guaranteed as this is the default counter and it occasionally gets bumped by background operations. There's no easy way to prevent these from happening so these cases are removed from this suite. Resolves: cockroachdb#128523, cockroachdb#128640 Epic: None Release note: None Co-authored-by: Bergin Dedej <[email protected]> Co-authored-by: David Hartunian <[email protected]>
128857: changefeedccl: attempt to deflake distribution strategy unit tests r=rharding6373 a=andyyang890 This patch attempts to deflake the changefeed distribution strategy unit tests by replacing all the manually constructed `ALTER TABLE ... RELOCATE` statements with a single statement that uses a select query. Informs cockroachdb#125170 Informs cockroachdb#125231 Informs cockroachdb#127027 Informs cockroachdb#127313 Release note: None Co-authored-by: Andy Yang <[email protected]>
The buffer ingester could get stuck during closing because it would independently evaluate a `running` bool and then try to write to a channel with a buffer of 1. This channel could have accumulated a payload prior to the ingester shutting down and would then block forever. Instead of using the boolean, the flush operation now selects on a channel that we close on shutdown. This ensure that no matter how many extra flushes we have blocked, they will all get cancelled. Resolves: cockroachdb#128453 Epic: None Release note: None
128713: sqlstats: add close chan to buffer ingester r=xinhaoz a=dhartunian The buffer ingester could get stuck during closing because it would independently evaluate a `running` bool and then try to write to a channel with a buffer of 1. This channel could have accumulated a payload prior to the ingester shutting down and would then block forever. Instead of using the boolean, the flush operation now selects on a channel that we close on shutdown. This ensure that no matter how many extra flushes we have blocked, they will all get cancelled. Resolves: cockroachdb#128453 Epic: None Release note: None Co-authored-by: David Hartunian <[email protected]>
Fixes a log line. Previously, when a lease was acquired, we'd log > lease acquisition succeeded: <empty> This was because the printed field `status.Lease` came from a zero `status`. `status` would only be populated if no lease acquisition had been necessary in the first place. It now prints the lease. Epic: none Release note: None
The lease acquisition loop is too silent, as seen in a recent investigation[^1]. If we loop around multiple times, we ought to at least log the status in each case which will give a clue as to what was wrong with the previous lease. [^1]: https://cockroachlabs.slack.com/archives/C0KB9Q03D/p1723534154297719?thread_ts=1723502058.001399&cid=C0KB9Q03D Epic: none Release note: None
128447: github: update CODEOWNERS for product-security team r=souravcrl a=pritesh-lahoti This PR updates code ownership for the `product-security` team. This team is responsible for all the customer-facing security features (AuthN, AuthZ, Network Security, etc.). Release note: None Co-authored-by: Pritesh Lahoti <[email protected]>
Previously we didn't include the locality of the remote node when we dialed a node. This prevented us from capturing locality aware stats for the connections. Epic: CRDB-41138 Release note: none
This commit adds the nodes locality information into the ContextOptions. This allows metrics to consult this to determine if a connection is from a remote locality. Epic: CRDB-41138 Release note: None
Some of the places that call UnvalidatedDial have the locality. By passing it in when it is known they will more accurately update the statistics. Epic: CRDB-41138 Release note: None
Extract a constant to make it easier to change the expected count. Epic: CRDB-41138 Release note: None
129241: logictest: deflake cross_version_tenant_backup r=rafiss a=rafiss This makes the test less flaky by making sure finalization is complete before beginning the RESTORE. fixes cockroachdb#129066 Release note: None Co-authored-by: Rafi Shamim <[email protected]>
Update pkg/testutils/release/cockroach_releases.yaml with recent values. Epic: None Release note: None Release justification: test-only updates
129290: master: Update pkg/testutils/release/cockroach_releases.yaml r=rail a=cockroach-teamcity Update pkg/testutils/release/cockroach_releases.yaml with recent values. Epic: None Release note: None Release justification: test-only updates Co-authored-by: CRL Release bot <[email protected]>
129227: kvserver/rangefeed: move perRangeEventSink to rangefeed pacakge r=nvanbenschoten a=wenyihu6 **kvserver/rangefeed: change p.Register to use registration interface** Previously, we introduced a registration interface to abstract the implementation details of buffered and unbuffered registration. This patch updates the p.Register function to use this interface as well. Part of: cockroachdb#126560 Release note: none ---- **kvserver/rangefeed: move perRangeEventSink to rangefeed pacakge** Previously, perRangeEventSink was defined in pkg/server/node.go. This patch relocates it to the rangefeed package to facilitate future commits testing within the rangefeed package itself. Part of: cockroachdb#126560 Release note: none Co-authored-by: Wenyi Hu <[email protected]>
Currently, we are using spot VMs only to run benchmark tests. The reason for the low adoption of spot VM is because of the test failures due to VM preemption. But, with the recent changes, we have better control on VM preemption changes where if there is a test failure due to preemption, the test is run on an on-demand VM the next time. Also, the current failed tests that are run on spot VM is 148 out of 1093 = 13.5%. If I consider failure out of total tests run is 148 out of 2508 = 5.9%. So, this PR removes the condition to use spot VMs only for benchmark tests and changes the probability to 75%. Fixes: cockroachdb#127236 Epic: None
When a test overrides sql.catalog.descriptor_lease_duration to 0, then it was possible for the test to starve itself by creating a timer with zero duration that is repeatedly reset. The bug occurred since after refreshTimer expired, it would call the jitteredLeaseDuration helper to get the new duration for the timer. If a test had overriden sql.catalog.descriptor_lease_duration to 0, this would cause the timer to keep being reset to 0 and starve other goroutines. Release note: None
Previously, irrespective of the restart flag value we ended up overwriting the startup script which led to defaulting the values of other flags passed previously to the start command. This was inadequate because when restarting, i.e., `restart=true`, it makes more sense that we would reuse the existing startup script by default. This patch changes this behaviour. Epic: none Fixes: cockroachdb#125085 Release note: None
128898: roachprod: reuse start script if start is called with --restart=true r=herkolategan,renatolabs a=vidit-bhat Previously, irrespective of the restart flag value we ended up overwriting the startup script which led to defaulting the values of other flags passed previously to the start command. This was inadequate because when restarting, i.e., `restart=true`, it makes more sense that we would reuse the existing startup script by default. This patch changes this behaviour. Epic: none Fixes: cockroachdb#125085 Release note: None Co-authored-by: Vidit Bhat <[email protected]>
127896: roachtest: use spot VM for all tests r=shailendra-patel,vidit-bhat a=nameisbhaskar Currently, we are using spot VMs only to run benchmark tests. The reason for the low adoption of spot VM is because of the test failures due to VM preemption. But, with the recent changes, we have better control on VM preemption changes where if there is a test failure due to preemption, the test is run on an on-demand VM the next time. Also, the current failed tests that are run on spot VM is 148 out of 1093 = 13.5%. If I consider failure out of total tests run is 148 out of 2508 = 5.9%. So, this PR removes the condition to use spot VMs only for benchmark tests and changes the probability to 75%. Fixes: cockroachdb#127236 Epic: None Co-authored-by: Bhaskarjyoti Bora <[email protected]>
129001: ui: display very large ranges in problematic ranges r=iskettaneh a=iskettaneh This commit adds a column in the DB Console's page problematic ranges that displays the ranges that are too large. The threshold is set to be: 8 * the current range max size. The two screenshots below show how that looks like. In order to generate that, I had to set the threshold multiplier to be 0.8 (instead of 8). <img width="600" alt="Screenshot 2024-08-15 at 11 13 28 AM" src="https://github.com/user-attachments/assets/66aae63a-1140-4079-a12e-b925110d975e"> <img width="600" alt="Screenshot 2024-08-15 at 11 14 15 AM" src="https://github.com/user-attachments/assets/fb7e9f07-1fd6-473c-bdd6-b902af3c3306"> Fixes: cockroachdb#127843 Epic: None Release note (ui change): Adding too large ranges to the problematic ranges page in DB Console. If a range is larger than twice the max range size, it will show up in the problematic ranges page. Co-authored-by: Ibrahim Kettaneh <[email protected]>
129085: roachtest: multi-region/mixed-version support for shared-process r=srosenberg,herkolategan a=renatolabs We enable required features on tenants before setting up the TPCC workload on the cluster. Informs: cockroachdb#127378 Release note: None Co-authored-by: Renato Costa <[email protected]>
123848: jobs: avoid CTEs in crdb_internal.system_jobs query r=dt a=stevendanna The CTE in the query used for crdb_internal.system_jobs can prevent a number of useful query optimizations. Informs cockroachdb#122687 Release note (performance improvement): Further improves the performance of job-system related queries. 129117: roachtest: update rebalance-load mixed-version tests for shared-process r=srosenberg a=renatolabs In this commit, we update the `rebalance/by-load` mixed-version tests so that they can run on shared-process deployments. As usual, we need to enable some features on tenants before initializing the workload. In addition, we pass the virtual cluster cookie when fetching system metrics. Informs: cockroachdb#127378 Release note: None 129302: lease: fix goroutine starvation caused by test cluster setting r=rafiss a=rafiss When a test overrides sql.catalog.descriptor_lease_duration to 0, then it was possible for the test to starve itself by creating a timer with zero duration that is repeatedly reset. The bug occurred since after refreshTimer expired, it would call the jitteredLeaseDuration helper to get the new duration for the timer. If a test had overriden sql.catalog.descriptor_lease_duration to 0, this would cause the timer to keep being reset to 0 and starve other goroutines. fixes cockroachdb#129097 fixes cockroachdb#129268 Release note: None Co-authored-by: Steven Danna <[email protected]> Co-authored-by: Renato Costa <[email protected]> Co-authored-by: Rafi Shamim <[email protected]>
Changes: * [`3d14906a`](cockroachdb/pebble@3d14906a) wal: fix reader buffer reuse * [`c2749cc7`](cockroachdb/pebble@c2749cc7) colblk: use crlib/crbytes.CommonPrefix * [`f90b350e`](cockroachdb/pebble@f90b350e) docs/rfc: add RFC for point tombstone density compaction heuristic * [`1a5295f8`](cockroachdb/pebble@1a5295f8) block: move block handle encoding and decoding * [`0e8a11a5`](cockroachdb/pebble@0e8a11a5) metamorphic: no synthetic suffix with overlapping RangeKeySets Release note (bug fix): Fix bug in the public preview WAL failover feature that could prevent a node from starting if it crashed during a failover. Epic: none.
This commit fixes a comment in the QueryTxnResponse.QueriedTxn field. It also addresses a related TODO. Epic: None Release note: None
Implementation of `SHOW LOGICAL REPLICATION JOBS`. An LDR job description column will be added in a follow-up PR. Rows returned by `SHOW LOGICAL REPLICATION JOBS` should include the following columns: - job_id: ID of an LDR job - status: job status - targets: names of tables being replicated in an LDR job - replicated_time: ingestion frontier timestamp An optional clause `WITH DETAILS` can be included in the query. If `WITH DETAILS` option is included, there should be an additional column in the rows returned by the query: - replication_start_time: initial timestamp from which the replication producer job will begin streaming MVCC revisions Epic: CRDB-40869 Release note: None
This commit does two things to deflake the test: 1) Takes the MinExpiration into consideration when advancing the clock to cause the epoch lease to expire. This is important after the feature was enabled in: ba758c0 2) Waits for N1's view of N2's lease expiration to match N2's view. This is important in the rare case where N1 tries to increase N2's epoch, but it has a stale view of the lease expiration time. Fixes: cockroachdb#124178, cockroachdb#123551 Epic: None Release note: None
129279: kvserver: deflake TestRangefeedCheckpointsRecoverFromLeaseExpiration r=iskettaneh a=iskettaneh This commit does two things to deflake the test: 1) Takes the MinExpiration into consideration when advancing the clock to cause the epoch lease to expire. This is important after the feature was enabled in: ba758c0 2) Waits for N1's view of N2's lease expiration to match N2's view. This is important in the rare case where N1 tries to increase N2's epoch, but it has a stale view of the lease expiration time. Fixes: cockroachdb#124178, cockroachdb#123551 Epic: None Release note: None Co-authored-by: Ibrahim Kettaneh <[email protected]>
129331: go.mod: bump Pebble to 3d14906a0e0c r=itsbilal a=jbowens Changes: * [`3d14906a`](cockroachdb/pebble@3d14906a) wal: fix reader buffer reuse * [`c2749cc7`](cockroachdb/pebble@c2749cc7) colblk: use crlib/crbytes.CommonPrefix * [`f90b350e`](cockroachdb/pebble@f90b350e) docs/rfc: add RFC for point tombstone density compaction heuristic * [`1a5295f8`](cockroachdb/pebble@1a5295f8) block: move block handle encoding and decoding * [`0e8a11a5`](cockroachdb/pebble@0e8a11a5) metamorphic: no synthetic suffix with overlapping RangeKeySets Release note (bug fix): Fix bug in the public preview WAL failover feature that could prevent a node from starting if it crashed during a failover. Epic: none. Informs cockroachdb/pebble#3865. Co-authored-by: Jackson Owens <[email protected]>
129333: kv: fix comment on QueryTxnResponse.QueriedTxn r=nvanbenschoten a=nvanbenschoten This commit fixes a comment in the QueryTxnResponse.QueriedTxn field. It also addresses a related TODO. Epic: None Release note: None Co-authored-by: Nathan VanBenschoten <[email protected]>
129095: crosscluster, sql: implement SHOW LOGICAL REPLICATION JOBS r=msbutler a=azhu-crl Implementation of `SHOW LOGICAL REPLICATION JOBS`. An LDR job description column will be added in a follow-up PR. Rows returned by `SHOW LOGICAL REPLICATION JOBS` should include the following columns: - job_id: ID of an LDR job - status: job status - targets: names of tables being replicated in an LDR job - replicated_time: ingestion frontier timestamp An optional clause `WITH DETAILS` can be included in the query. If `WITH DETAILS` option is included, there should be an additional column in the rows returned by the query: - replication_start_time: initial timestamp from which the replication producer job will begin streaming MVCC revisions Example output of `SHOW LOGICAL REPLICATION JOBS`: ``` [email protected]:26257/demoapp/b> SHOW LOGICAL REPLICATION JOBS; job_id | status | targets | replicated_time ---------------------+---------+---------+------------------------- 996466532408918017 | running | {tab} | 2024-08-20 15:32:30+00 996466532446535681 | running | {tab} | 2024-08-20 15:32:30+00 (2 rows) Time: 14ms total (execution 14ms / network 1ms) ``` Example output of `SHOW LOGICAL REPLICATION JOBS WITH DETAILS`: ``` [email protected]:26257/demoapp/b> SHOW LOGICAL REPLICATION JOBS WITH DETAILS; job_id | status | targets | replicated_time | replication_start_time ---------------------+---------+---------+------------------------+-------------------------------- 996466532408918017 | running | {tab} | 2024-08-20 15:32:30+00 | 2024-08-20 15:30:52.514488+00 996466532446535681 | running | {tab} | 2024-08-20 15:32:30+00 | 2024-08-20 15:30:52.526056+00 (2 rows) Time: 10ms total (execution 10ms / network 1ms) ``` Epic: [CRDB-40869](https://cockroachlabs.atlassian.net/browse/CRDB-40869) Release note: None Co-authored-by: Anne Zhu <[email protected]>
Release note: None
129335: changefeedccl: add more logging around core changefeed completion r=rharding6373 a=andyyang890 Epic: CRDB-37337 Release note: None Co-authored-by: Andy Yang <[email protected]>
129149: ui: remove unnecessary metric name prefixes r=aadityasondhi a=dhartunian Metrics keyed by nodeID and prefixed with long strings become very difficult to read on the dashboard pages since they take up too much horizontal space. In all of these cases, the context provided by the graph title and Y axis label provide enough context to omit verbose legend labels. Resolves: cockroachdb#128046 Release note (ui change): some metric charts on the overview and replication pages have more terse legends to facilitate easier browsing. Co-authored-by: David Hartunian <[email protected]>
129058: roachprod: azure GC supports cleanup for multiple subscriptions r=srosenberg a=DarrylWong Previously Azure GC assumed that only one subscription ID would be used at a time. However the current structure of our Azure subscriptions makes a distinction between user "adhoc" roachprod clusters and nightly "infra" clusters. As a result, roachprod clusters can live in two different subscriptions. This change adds the azure-subscription-name flag to support cleaning up in multiple subscriptions. Additionally it also switches the default subscription ID to e2e-adhoc, which is what users should use. Fixes: cockroachdb#129057 Release note: none Epic: none Co-authored-by: DarrylWong <[email protected]>
Epic: none Release note: None
129368: kvclient: improve logging in OptimizeReplicaOrder r=kvoli a=arulajmani Epic: none Release note: None Co-authored-by: Arul Ajmani <[email protected]>
129131: changefeedccl: enable vmodule for TestAlterChangefeedAddTargetFamily r=rharding6373 a=andyyang890 This patch enables verbose logging (vmodule `helpers_test=1`) for `TestAlterChangefeedAddTargetFamily` so that we can get more information about what, if any, messages that the changefeed sees before the test times out. Informs cockroachdb#128749 Release note: None Co-authored-by: Andy Yang <[email protected]>
This change disbales the CI step for pull requests but this still runs for push step into the target branch. Also the count of the microbenchmarks are increased to get more robust results. Epic: none Release note: None
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.