Zero autoinc mode #6749
Zero autoinc mode #6749
Conversation
This diff introduces -seq-auto-value-on-zero as a vtgate flag. Until now sequences have only populated NULL columns, which is not consistent with the default behavior of mysql. It's consistent with the NO_AUTO_VALUE_ON_ZERO sql mode. In a perfect world we might want to detect per-keyspace sql modes. For now, it seems simple and adequate to let people select at the vtgate level if they want this behavior or not for all keyspace. The motivation for this diff is that we have some situations where people still test code against direct mysql without vtgate. We've twice had folks commit insert statements with 0 for a sequence column that worked in a local test but failed against vtgate. Aligning behavior of vtgate with mysql seems like the cleanest way to prevent that in the future. Signed-off-by: David Weitzman <dweitzman@pinterest.com>
Signed-off-by: David Weitzman <dweitzman@pinterest.com>
Signed-off-by: Derrick Laird <swampdonk@gmail.com>
353b0fd to
8ac2868
Compare
|
What's the thinking behind removing the option to opt-out of this and changing the behavior by default? Backwards-incompatible changes seems to carry some risk, particularly with no option to opt into the old behavior. Seems like to do this by default without a flag there would have to be general consensus that this is a bug fix and not a behavior change. To me it's not obvious that this is a bug fix (vs a behavior change). The stated policy for backwards-incompatible flag default value changes is that they flip during major version releases: https://github.com/vitessio/enhancements/blob/master/veps/vep-1.md#backwards-and-forwards-compatibility-promise Seems like introducing a flag that's off by default in one major release and toggling the default in a future major release is probably the way to go?... |
|
Looks like @dweitzman attempted this some time ago #4751. I voted for changing the behavior (without a flag) because it's the mysql default behavior. This is not likely to break anyone. I suggest that we just announce it in a RFC, which will make it to the release notes. |
Works for me. For our use case this is definitely an improvement in behavior |
|
@MordFustang21 can you create an issue that announces this as a breaking change? Once that's done, we can merge this PR. Here's an example #6065 |
This adds support to generate a sequence value when a 0 is passed instead of just null. This is the default behavior for MySQL.
Also need to look into detecting when sql_mode has been set to NO_AUTO_VALUE_ON_ZERO and handle that accordingly. I'm not sure the best way to go about that so any guidance there would be appreciated.
Ref #4751