In domain mode, the "restart-servers" command after application deployment in batch fails with the following message.
$ cat deploy.cli batch deploy /tmp/kitchensink.war --server-groups=production-group :restart-servers(blocking=true) run-batch $ ./bin/jboss-cli.sh -c --file=/path/to/deploy.cli The batch failed with the following error (you are remaining in the batch editing mode to have a chance to correct the error): {"WFLYDC0074: Operation failed or was rolled back on all servers. Server failures:" => {"server-group" => {"production-group" => {"host" => {"host2" => {"production-server-A" => {"WFLYCTL0062: Composite operation failed and was rolled back. Steps that failed:" => {"Operation step-1" => {"WFLYCTL0062: Composite operation failed and was rolled back. Steps that failed:" => {"Operation step-1" => {"WFLYCTL0062: Composite operation failed and was rolled back. Steps that failed:" => {"Operation step-1" => "WFLYCTL0212: Duplicate resource [(\"deployment\" => \"kitchensink.war\")]"}}}}}}},"host3" => {"production-server-B" => {"WFLYCTL0062: Composite operation failed and was rolled back. Steps that failed:" => {"Operation step-1" => {"WFLYCTL0062: Composite operation failed and was rolled back. Steps that failed:" => {"Operation step-1" => {"WFLYCTL0062: Composite operation failed and was rolled back. Steps that failed:" => {"Operation step-1" => "WFLYCTL0212: Duplicate resource [(\"deployment\" => \"kitchensink.war\")]"}}}}}}}}}}}}
Also, there remains directories in data/content on servers. These directories need to be deleted by OS commands.
$ tree ./machine2/servers/production-server-A/data/content/ ./machine2/servers/production-server-A/data/content/ └── b5 └── 4386a025e5bca365d46ebcc2b2714e69ded255 └── content3 directories, 1 file $ tree ./machine3/servers/production-server-B/data/content/ ./machine3/servers/production-server-B/data/content/ └── b5 └── 4386a025e5bca365d46ebcc2b2714e69ded255 └── content3 directories, 1 file
After deleting these directories, it seems that the servers needs to be restarted. It seems that EAP is trying to rollback only domain controller and the content repository after failing application deployment. It is necessary for host controllers and servers to be rolled back (including data/content).
If "restart-servers" command is not used in batch, it succeeds.
$ cat deploy2.cli batch deploy /tmp/kitchensink.war --server-groups=production-group /host=host2/server=production-server-A:restart /host=host3/server=production-server-B:restart run-batch $ ./bin/jboss-cli.sh -c --file=/path/to/deploy2.cli The batch executed successfully
Please check "readme.txt" in reproducer.zip.
- clones
-
JBEAP-28080 [GSS](7.4.z) "restart-servers" fails after application deployment in batch and content directories remains on domain mode
-
- New
-
