I have 3 node cluster, problem is that for some rows , updatetime column have different value on 3 node: like : 1st node : updated_on | 2018-08-22 17:01:30+0530 2nd node : updated_on | 2018-08-22 17:02:30+0530 3rd node : updated_on | 2018-08-22 17:03:30+0530

Please suggest , where i need to digging.

  • show the schema for the table, and what query is executed
    – Alex Ott
    Aug 22 '18 at 18:28
  • CREATE TABLE prod.ufi( file_id bigint, filename text, created_by text, created_on timestamp, file_process_status text, guid text, m_info map<text, text>, metadata text, mid text, status text, uid text, updated_by text, updated_on timestamp, wallet_info map<text, text>, PRIMARY KEY (file_id, filename) ) ; Simply just update the row through application.
    – AmitPatel
    Aug 23 '18 at 7:41
  • how often this update happens - is it possible that there were 3 updates during the 3 minutes? What consistency level is used for updating - that standard LOCAL_ONE?
    – Alex Ott
    Aug 23 '18 at 7:50
  • consistency level is ONE. This issue happens sometimes not frequently. But, if 3 updates are running in 3 different time, then last update upsert rest 2.
    – AmitPatel
    Aug 23 '18 at 11:31

Depending on the version of the driver you're using, the timestamp of the data could be generated either on server, or on client (Java driver 3.x, for example). You need to check that all servers and client machines are synchronized via ntp.

  • Yes all app and cassandra node are sync via NTP
    – AmitPatel
    Aug 23 '18 at 13:33
  • Ok, it could be that some nodes don't receive updates if they are down, for example, if hints are disabled. Try to perform repairs in your cluster
    – Alex Ott
    Aug 24 '18 at 14:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.