Through ops-center and nodetool cfstats i was able to find that one of the partitions of a keyspace table is 560 Mb, but couldn't find out which partition is that. How can we trace which partition of the table is that big ??
How to trace back a large partition of a column family in cassandra
Asked Answered
The fastest possible way is to look for messages in the log about compacting large partitions. Sort of a cheat, but it often works.
Short of that, you'll need to dump the sstables to json, and then inspect the json. There are a number of people who have written tools for this online - https://github.com/BrianGallew/cassandra_tools is one example.
Hai @ Jeff Jirsa can i try writing shell script which can scan all sstables on the node and find the partition, but the thing i would like to know is how can we find how partitions are separated in sstables and how can we determine size probably by counting no.of lines or characters though i think i cannot find it exactly partitions but i think i can find couple of big partitions which as later search. –
Iaea
cat /var/log/cassandra/system.log | grep large
did the trick. Awesome, thank you! –
Annaleeannaliese © 2022 - 2024 — McMap. All rights reserved.