Open BalaMahesh opened 2 years ago
@findepi any idea why this is happening , it's strange no else has faced this issue earlier ?
i do not, but i'd ask @electrum for advice.
@electrum can you please take a look and suggest on this.
@findepi ,
@electrum seems busy.
can you please add any others who can help with this.
@electrum can you please guide on this.
@findepi @electrum I am also facing the same issue. Can you please help here?
Additional, relevant slack thread https://trinodb.slack.com/archives/CGB0QHWSW/p1653301299029359
trino version : 366 Issue : We are running INSERT query which writes data into the bucketed hive table in orc format. In hive connector we have set the below config
hive.temporary-staging-directory-enabled = true hive.temporary-staging-directory-path = /data0/logs/trino (changed from /tmp/presto-${USER} , since we are facing this and wanted to try with different path).
Case 1: With 70 workers in the cluster, this query is able to finish successfully and writes data into hive table in orc format. In this case we could see the tmp files created in the worker nodes and once the query is finished, temp files get deleted. Attached the screenshot of temp files with names from one of the worker nodes.
Case 2: With 60 workers in the cluster and the same hive connector configuration, the query is failing with the below error from one of the worker nodes. I could see the tmp files created for the query in other worker nodes(ssh ing to all the worker nodes at the same time is not easy) and they are deleted after the query is failed.
Can someone suggest the possible reason and fix for this issue. Thanks in advance.