importccl: Modify row_id generation for IMPORT from CSV. #39148
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
|
This change is |
dt
approved these changes
Jul 30, 2019
This change was motivated by a bug in which consecutive IMPORT INTO queries into a table without an explicit PK, and from unique data sources would overwrite instead of appending data. This is because row_id generation was based on fileIndex and rowNum which created colliding PKs across query runs. To fix this we add the timestamp at which IMPORT INTO is run at a 10-microsecond granularity to the rowNum. Release note: None
|
bors try |
tryBuild succeeded |
|
TFTR |
craig bot
pushed a commit
that referenced
this pull request
Jul 30, 2019
39148: importccl: Modify row_id generation for IMPORT from CSV. r=adityamaru27 a=adityamaru27 This change was motivated by a bug in which consecutive IMPORT INTO queries into a table without an explicit PK, and from unique data sources would overwrite instead of appending data. This is because row_id generation was based on fileIndex and rowNum which created colliding PKs across query runs. To fix this we add the timestamp at which IMPORT INTO is run at a 10-microsecond granularity to the rowNum. 39166: server: fix the advisory certificate check r=knz a=knz Fixes #32539. (context: `cockroach start` performs some advisory checks on the certificates prior to starting up, to inform the user upfront if the cert configuration is suspicious. These checks are complementary to the actual security validation performed in the go std lib.) Prior to this patch, the advisory cert validation was performing checks on both the addres given to `--listen` and that given to `--advertise` (either explicitly, or implicitly when derived from `--listen`). This was problematic because the listen address is really irrelevant to security validation. What matters for node cert validation is that the advertised address is present. So this patch removes the check on the listen addr (recommended by @mberhault). Additionally this patch also enhances the check on the advertise addr: if a host name is given to `--advertise-addr` (either explicitly, or implicitly as derived from `--listen-addr`), and the name is not present in the cert, the validation is considered to pass if the resolved address for the given name is present int he cert. Release note (cli change): The advisory/informative check performed by `cockroach start` on the validity of addresses contained in the node certificate is relaxed to focus on the advertised node address, and to tolerate cases when the cert contains an IP address but a hostname is specified as advertised address. Co-authored-by: Aditya Maru <adityamaru@cockroachlabs.com> Co-authored-by: Raphael 'kena' Poss <knz@cockroachlabs.com>
Build succeeded |
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.
This change was motivated by a bug in which consecutive IMPORT INTO
queries into a table without an explicit PK, and from unique data
sources would overwrite instead of appending data.
This is because row_id generation was based on fileIndex and
rowNum which created colliding PKs across query runs.
To fix this we add the timestamp at which IMPORT INTO is run
at a 10-microsecond granularity to the rowNum.