You are correct. Thank you! One of the machines in the cluster is a
Centos 8 box that doesn't come with NTP - they only included
chronyd...which doesn't seem to work with our existing NTP server.
Ugh, I finally gave up and just set the clock manually and it now works.
Cheers!
-Joe
On
I believe you could set your tables to flush to disk at specific intervals
(memtable_flush_period_in_ms), note that you'd have to set this for all
tables (not just the CDC enabled tables) to ensure that commitlog files are
flushed to the cdc_raw directory. Or as Dhanunjaya noted you could just
Want to add - I am seeing this in the log:
INFO [ScheduledTasks:1] 2021-05-05 09:36:05,022
MessagingMetrics.java:206 - COUNTER_MUTATION_RSP messages were dropped
in last 5000 ms: 0 internal and 1 cross node. Mean internal dropped
latency: 0 ms and Mean cross-node dropped latency: 21356 ms
I'm seeing some odd behavior with RC1 and counters - from cqlsh:
cqlsh> select * from doc.seq;
id | doccount
--+--
DS | 1
DS_1 | 844
(2 rows)
cqlsh> update doc.seq set doccount=doccount+1 where id='DS_1';
OperationTimedOut: errors={'172.16.100.208:9042': 'Client
This sounds like the clock on your Cassandra servers are not in sync.
Can you please ensure all Cassandra servers have their clock synced
(usually via NTP) and retry this?
On 05/05/2021 14:42, Joe Obernberger wrote:
Want to add - I am seeing this in the log:
INFO [ScheduledTasks:1]
NTP is running and clocks are sync'd?
No large pauses?
On Wed, May 5, 2021 at 6:44 AM Joe Obernberger
wrote:
> Want to add - I am seeing this in the log:
> INFO [ScheduledTasks:1] 2021-05-05 09:36:05,022
> MessagingMetrics.java:206 - COUNTER_MUTATION_RSP messages were dropped
> in last 5000
