Benoit Tellier created JAMES-3886:
-------------------------------------
Summary: LocalDelivery failure when mailbox counter update
Key: JAMES-3886
URL: https://issues.apache.org/jira/browse/JAMES-3886
Project: James Server
Issue Type: New Feature
Components: cassandra, mailbox
Affects Versions: 3.7.3
Reporter: Benoit Tellier
Fix For: 3.7.4, 3.8.0
As part of a customer support contract, we noticed that a failure to update
mailbox counters results in a delivery failures.
Log line:
{code:java}
2023-02-10T10:10:57,537 [WARN ] o.a.j.t.m.d.MailDispatcher - Error While
storing mail. This error will be retried.
com.datastax.driver.core.exceptions.WriteTimeoutException: Cassandra timeout
during COUNTER write query at consistency ONE (1 replica were required but only
0 acknowledged the write)
{code}
Consistency level ONE usage is inherent to counter usage. Thus errors are to be
expected in case of single Cassandra node failure.
I hereby proposes to IGNORE failure to update counters and just fire a WARNING
log to inform the administrator.
Rationnals:
- The mail is already stored in a mailbox
- counters can be seen as not business critical but a mere indicator
- And we have the tools (applicative read repairs + corrective tasks) to fix
the inconsistencies CF
https://github.com/apache/james-project/blob/master/src/adr/0023-cassandra-mailbox-counters-inconsistencies.md
https://github.com/apache/james-project/blob/master/src/adr/0042-applicative-read-repairs.md
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
