The problem was that we were expecting last raft log record to be very
specific one but vshard rebalancing came and added some dml record. Use
pending_schema_change from pico property instead.

Close: https://git.picodata.io/picodata/picodata/picodata/-/issues/554

Arseniy Volynets authored 8 months ago and Dmitry Rodionov committed 8 months ago

- we didn't wait until replication catches up before
checking dml results on other instance. Fix this
by doing quorum read on the other instance

An element of array 'nc->stack' of size 256, declared at
lj_opt_narrow.c:367, is accessed by an index with possible value -1 at
lj_opt_narrow.c:377. This may lead to a buffer overflow.

After having been assigned to a NULL value at printf.c:219, pointer
'bufpt' is passed as 2nd parameter in call to function
'sqlStrAccumAppend' at printf.c:798, where it is dereferenced at
printf.c:903.

Warning:
Return value of a function 'space_cache_find' is dereferenced at
trigger.c:364 without checking for NULL, but it is usually checked for
this function (33/34).

Follow up for 33dc7181. Tests didnt
catch that.

Follow up for c6d684dd

- update sbroad submodule to commit with helper
patches for dml on global tables
- add support for sql dml insert/update/delete
on global tables. Insert on conflict is not
supported yet
- Current implementation uses a single batch of
dml commands via CAS. No additional predicates
are used in CAS, so no isolation level is guaranteed
even when query only refers to global tables

1. make it non-nullable
2. fix tests

- update sbroad submodule to commit supporting
inferring not null constraint on primary key
columns
- remove redundant not null constraint from
most integration tests

Kurdakov Alexander authored 9 months ago and Alexander Kurdakov committed 8 months ago

following sections were renamed:
tiers -> tier
peers -> peer