Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Browse files
Browse the repository at this point in the history
61406: sql: validate against array type usages when dropping enum values r=ajwerner a=arulajmani Previously, when dropping an enum value, we weren't validating if the enum value was used in a column of array type. This patch fixes the bug. Fixes #60004 Release justification: bug fix to new functionality Release note: None 61650: roachtest: exclude `lost+found` directory r=knz a=rickystewart This directory is created by the filesystem and unowned file chunks are put there by `fsck`. The directory and its contents aren't readable by anyone except `root`, so this can cause the `du -c /mnt/data1` that `roachtest` performs to fail -- add an `--exclude` to handle this. We already ignore this directory in other contexts (for example, see `pkg/storage/mvcc.go`). Fixes #53663. Release justification: Non-production code change Release note: None 61788: importccl: unskip userfile benchmark r=pbardea a=adityamaru I've run this ~20 times and it averages ~13s to run. I suspect the fixes to linked issues mentioned in #59126 might have mitigated this. Going to unskip due to lack of reproducibility. Fixes: #59126 Release note: None 61828: contention: store contention events on non-SQL keys r=yuzefovich a=yuzefovich Previously, whenever we tried to add a contention event on a non-SQL key, it would encounter an error during decoding tableID/indexID pair, and the event was dropped. This commit extends the contention registry to additionally store information about contention on non-SQL keys. That information is stored in two levels: - on the top level, all `SingleNonSQLKeyContention` objects are ordered by their keys - on the bottom level, all `SingleTxnContention` objects are ordered by the number of times that transaction was observed to contend with other transactions. `SingleTxnContention` protobuf message is moved out of `SingleKeyContention` and is reused for non-SQL keys. This commit also updates the status server API response. I assume that no changes are needed with regards to backwards compatibility since the original version was merged just a few weeks ago, and we haven't had a beta released since then. Fixes: #60669. Release note (sql change): CockroachDB now also stores the information about contention on non-SQL keys. 61862: cliccl: add `load show backups` to display backup collection r=pbardea a=Elliebababa Previously, users can list backups created by `BACKUP INTO` with `SHOW BACKUP IN`in a sql session. But this listing task can be also done offline without a running cluster. This PR updates `load show` with `backups` subcommand, which allows users to list backups in a backup collection created by `BACKUP INTO`. With the same purpose as other `load show` subcommands, this update allows users to list backups without running `SHOW BACKUP IN` in a sql session. see #61131 #61829 to checkout other `load show` subcommand. Release note (cli change): Add `load show backups` to display backup collection. Previously, users can list backups created by `BACKUP INTO` via `SHOW BACKUP IN`in a sql session. But this listing task can be also done offline without a running cluster. Now, users are able to list backups in a collection with `cockroach load show backups <collection_url>. 61877: bench/ddl_analysis: fix test for real r=ajwerner a=ajwerner See individual commits. Last is the critical one. Fixes #61856. 61937: colexec: make vectorized stats concurrency safe r=yuzefovich a=yuzefovich **colflow: clean up vectorized stats for rowexec processors** Previously, the wrapped row-execution KV reading processors were implementing `execinfra.KVReader` interface, but they were never used as such, only the ColBatchScans would get used to retrieve the KV stats. This is the case because the row-execution processors report their execution stats themselves, and we don't want to duplicate that info. This commit moves `KVReader` interface into `colexecop` package and now only the ColBatchScans implement it. This allowed for some cleanup around the vectorized stats code, but the main reason for performing this change is that the contract of the interface will be modified by the follow-up commit to mention the safety under concurrent usage, and I didn't want to change the row-execution processors for that since the relevant methods never get called anyway. Additionally, this commit begins emitting of rows-read and bytes-read by the zigzagJoiners and invertedJoiners to complete the metrics picture. Release note: None **colexec: make vectorized stats concurrency safe** Previously, the collection of vectorized stats was not synchronized with the operators themselves. Namely, it was possible to call methods like `GetBytesRead` on the ColBatchScans and Inboxes from a different goroutine (the root materializer or the outbox) than from the main goroutine of the operator. This is now fixed by putting mutexes in place and updating `colexecop.KVReader` interface to require concurrency-safe implementations. Fixes: #61899. Release note: None Co-authored-by: arulajmani <arulajmani@gmail.com> Co-authored-by: Ricky Stewart <ricky@cockroachlabs.com> Co-authored-by: Aditya Maru <adityamaru@gmail.com> Co-authored-by: Yahor Yuzefovich <yahor@cockroachlabs.com> Co-authored-by: elliebababa <ellie24.huang@gmail.com> Co-authored-by: Andrew Werner <awerner32@gmail.com>
- Loading branch information
