> When a zedstore table is queried using *invalid* ctid, the server > crashes due to assertion failure. See below, > > postgres=# select * from t2 where ctid = '(0, 0)'; > server closed the connection unexpectedly > This probably means the server terminated abnormally > before or while processing the request. > The connection to the server was lost. Attempting reset: Failed.
Thank you for pointing that out! I will look into fixing that some time this week. If we run without assertions the query still fails with this error because zedstoream_tuple_tid_valid incorrectly reports the TID as valid: ERROR: arrived at incorrect block 2 while descending zedstore btree > I believe above should have either returned 1 rows or failed with some > user friendly error. Agreed. I think it should match the behavior of heap as closely as possible.