Re: null values injected while drop compact storage was executed
This happened with version 3.11.10. We were analyzing impact until now as this happened only on our production systems. There was probably not enough concurrency on our staging environments so it did not happen there. We will start to write a reproducer and file an issue afterwards. Thanks for the feedback Jeff and Scott! Best, Matthias
Re: null values injected while drop compact storage was executed
If you don't have an explicit goal of dropping compact storage, it's not necessary to as a prerequisite to upgrading to 4.x+. Development community members recognized that introducing mandatory schema changes as a prerequisite to upgrading to 4.x would increase operator + user overhead and limit adoption of the release. To address this, support for compact storage was reintroduced into 4.x which eliminated the requirement to drop it: https://issues.apache.org/jira/browse/CASSANDRA-16217 @Matthias, would you be willing to share the Apache Cassandra version you were running when you observed this behavior and to file a Jira ticket? – Scott On May 7, 2024, at 7:18 AM, Jeff Jirsa wrote: This sounds a lot like cassandra-13004 which was fixed, but broke data being read-repaired during an alter statement I suspect it’s not actually that same bug, but may be close/related. Reproducing it reliably would be a huge help. - Jeff On May 7, 2024, at 1:50 AM, Matthias Pfau via user wrote: Hi there, we just ran drop compact storage in order to prepare for the upgrade to version 4. We observed that column values have been written as null, if they where inserted while the drop compact storage statement was running. This just happened for the couple seconds the drop compact storage statement ran. Did anyone else observe this? What are the proposed strategies to prevent data loss. Best, Matthias
Re: null values injected while drop compact storage was executed
This sounds a lot like cassandra-13004 which was fixed, but broke data being read-repaired during an alter statement I suspect it’s not actually that same bug, but may be close/related. Reproducing it reliably would be a huge help. - Jeff > On May 7, 2024, at 1:50 AM, Matthias Pfau via user > wrote: > > Hi there, > we just ran drop compact storage in order to prepare for the upgrade to > version 4. > > We observed that column values have been written as null, if they where > inserted while the drop compact storage statement was running. This just > happened for the couple seconds the drop compact storage statement ran. > > Did anyone else observe this? What are the proposed strategies to prevent > data loss. > > Best, > Matthias
null values injected while drop compact storage was executed
Hi there, we just ran drop compact storage in order to prepare for the upgrade to version 4. We observed that column values have been written as null, if they where inserted while the drop compact storage statement was running. This just happened for the couple seconds the drop compact storage statement ran. Did anyone else observe this? What are the proposed strategies to prevent data loss. Best, Matthias