All geek questions in one place

Cannot find sufficient sources for streaming

4 nodes with 2 seeds were combined. Using cassandra 1.2.5 . The idea of ​​creating 2 clusters with 2 nodes with 1 seed in each cluster.

Discontinued, changed seeds in the cassandra.yaml file and deleted the commitlog data and directories and restarted, one node started normally, and the second for both clusters.

 INFO 10:28:37,241 JOINING: sleeping 30000 ms for pending range setup INFO 10:29:07,242 JOINING: Starting to bootstrap... ERROR 10:29:07,324 Exception encountered during startup

java.lang.IllegalStateException: Cannot find sufficient sources for (5253604194908817102,5256320442405919104]

+4
source share
2 answers

If all your data has been cleared and you are setting up new nodes, you should get this exception:

java.lang.IllegalStateException: Cannot find sufficient sources for the streaming range (5253604194908817102,5256320442405919104]

The exception means that you do not have enough data to send over the required ranges, for example, if you have a replication coefficient of 1 for a specific key space in a 4 node cluster and you delete 2 nodes, you will not be able to stream certain ranges because they are lost with decommissioned nodes.

From what I understand, you are trying to do this:

 Original Cluster: Cluster A (x2 seeds) [node1, node2, node3, node4] New Cluster(s): Cluster A (x1 seed) [node1, node2] <-- no data Cluster B (x1 seed) [node1, node2] <-- no data

It means that you removed 1 seed from the source cluster and decommissioned 2 nodes. Then the 2 nodes that were decommissioned were added to their own cluster (cluster B). And for all 4 nodes, you deleted the data / logging directories.

If you have done this, you should not transmit data streams. This means that you probably have not deleted all data directories. My advice is to use a clean C * installation and just configure cassandra.yaml for a new infrastructure of 2 clusters with 1 seed node each.

+5
source

Go to the /var/lib/cassandra/data/system directory and delete all subdirectories except %schema% . Use

 rm -rf batchlog compaction_history compactions_in_progress hints IndexInfo local NodeIdInfo paxos peer_events peers range_xfers sstable_activity

and restart your cassandra node.

0
source

Source: https://habr.com/ru/post/1494470/

More articles:


All Articles