Hello,
We are using Vertica 7.2 on a 4 nodes cluster, in a lab environment.
On the Vertica MC, we are observing this error quite frequently, in the context of the qualification of an SQL package.
We are more particularly concerned by one type of occurrence of this message, mentioning a join between 2 tables, one of them being always strictly empty.
In addition to that, we see that the initiator node of this error is always the same.
Finally, when we look at the results of all our queries, everything looks OK.
So I have several questions here:
1) Why would a join where one of the table is empty cause a memory problem?
2) Any reason why this is always the same node that is reporting the problem, and not the others ?
3) Should we really worry about the results of our query execution, given that we do not observe any symptoms other than this message in the console and in vertica logfile?
4) What can we change to prevent this error?
Thanks for helping,