Vadim Pakhnushev created IGNITE-26914:
-----------------------------------------
Summary: NPE in ClusterStateHttpServerFilter
Key: IGNITE-26914
URL: https://issues.apache.org/jira/browse/IGNITE-26914
Project: Ignite
Issue Type: Bug
Components: rest ai3
Reporter: Vadim Pakhnushev
Assignee: Vadim Pakhnushev
When node is restarted but REST requests are still coming, the following NPE
happens in the
{{ItCmgDisasterRecoveryTest.migratesManyNodesThatSawNoRepairToNewCluster}}:
{noformat}
[ERROR][default-nioEventLoopGroup-31-3][JavaExceptionHandler] Unhandled
exception
java.lang.NullPointerException: Cannot invoke
"org.apache.ignite.internal.rest.RestManager.pathAvailability(String)" because
"this.restManager" is null
at
org.apache.ignite.internal.rest.authentication.ClusterStateHttpServerFilter.doFilter(ClusterStateHttpServerFilter.java:51)
at
io.micronaut.http.filter.HttpServerFilter.doFilter(HttpServerFilter.java:48)
at
io.micronaut.http.server.RouteExecutor.filterPublisher(RouteExecutor.java:409)
at
io.micronaut.http.server.RouteExecutor.buildResultEmitter(RouteExecutor.java:634)
at
io.micronaut.http.server.RouteExecutor.executeRoute(RouteExecutor.java:354)
at
io.micronaut.http.server.netty.RoutingInBoundHandler.handleRouteMatch(RoutingInBoundHandler.java:594)
{noformat}
This is uncovered by the changes in IGNITE-26864 most likely due to different
timings of the incoming request, now it's much faster.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
