No that won't work. Failover works when the kafka servers are available at
application startup time and goes down sometime after the application is up and
running. That works fine for us. But if the kafka servers are not available
during the application startup time, the application fails to start giving an
exception. Availability/Unavailability of logging should not block an
application from starting up. The same issue is mentioned in the bug reference
I mentioned in my email.
Regards,Bhavesh.
On Monday, November 26, 2018, 10:31:36 AM PST, Matt Sicker
<[email protected]> wrote:
You can use the FailoverAppender as a workaround. Set up a file appender or
similar as your failover option so that logging still works. Adding a
feature to replay logs to an appender when it comes back online, however,
would be a separate task.
On Sun, 25 Nov 2018 at 23:21, Bhavesh Patel <[email protected]>
wrote:
> Hi Log4j2 developers, Do we have a
> status on bug (https://jira.apache.org/jira/browse/LOG4J2-2162)? Or is
> there a workaround for this issue? We are moving from log4j1 to log4j2 and
> will be using the Kafka appender for our production systems. This would be
> a major issue for us that if kafka server is down, the application itself
> fails to start up and would result in loss of revenues. IS there any
> recommendation to handle/bypass initializing the appender. We also have a
> rolling file appender which we can fallback to.
> Regards,Bhavesh.
>
>
>
--
Matt Sicker <[email protected]>
