Closed keith-turner closed 2 years ago
I have only tested this change w/ 2.1.0-SNAPSHOT and it fixed the problem I was seeing. I have not tried running Muchos w/ 2.0.0 though.
For completeness, can you share the details about the original error you encountered?
For completeness, can you share the details about the original error you encountered?
@arvindshmicrosoft this is the error I saw.
TASK [start accumulo 2.0 master, monitor, gc & tracer] **********************************************************************************************************************************************************
Wednesday 05 January 2022 17:53:37 +0000 (0:00:00.766) 0:03:01.386 *****
changed: [ktwal2-0] => (item=master)
changed: [ktwal2-0] => (item=monitor)
changed: [ktwal2-0] => (item=gc)
failed: [ktwal2-0] (item=tracer) => {"ansible_loop_var": "item", "changed": false, "cmd": ["nohup", "/home/centos/install/accumulo-2.1.0-SNAPSHOT/bin/accumulo-service", "tracer", "start"], "delta": "0:00:00.071548", "end": "2022-01-05 17:53:43.238652", "item": "tracer", "msg": "non-zero return code", "rc": 1, "start": "2022-01-05 17:53:43.167104", "stderr": "nohup: ignoring input\nUsage: accumulo-service <service> <command>\n\nServices:\n gc Accumulo garbage collector\n monitor Accumulo monitor\n manager Accumulo manager\n master Accumulo master (Deprecated)\n tserver Accumulo tserver\n compaction-coordinator Accumulo compaction coordinator (experimental)\n compactor Accumulo compactor (experimental)\n\nCommands:\n start Starts service\n stop Stops service\n kill Kills service", "stderr_lines": ["nohup: ignoring input", "Usage: accumulo-service <service> <command>", "", "Services:", " gc
Accumulo garbage collector", " monitor Accumulo monitor", " manager Accumulo manager", " master Accumulo master (Deprecated)", " tserver Accumulo tserver", " compaction-coordinator Accumulo compaction coordinator (experimental)", " compactor Accumulo compactor (experimental)", "", "Commands:", " start Starts service", " stop Stops service", " kill Kills service"], "stdout": "Invalid arguments: 'tracer' is an invalid <service>", "stdout_lines": ["Invalid arguments: 'tracer' is an invalid <service>"]}
I think the problem was caused by the tracer service being removed in apache/accumulo#2259
- ansible/wipe-systemd.yml: the accumulo-tracer service won't exist so it can't be stopped
It can be made conditional for sure though, but JFYI it is benign as Ansible will not error when the service does not exist.
- ansible/roles/accumulo/files/accumulo-cluster-systemd: this might need to be turned into a jinja template so the tracer sections can be included conditionally
This should be ideally addressed in this PR itself. Currently the accumulo-cluster-systemd
script errors out when run with 2.1.0-SNAPSHOT and systemd.
- ansible/wipe-systemd.yml: the accumulo-tracer service won't exist so it can't be stopped
Based on what @arvindshmicrosoft said, I think it would be best to do nothing for this case. Keeps the ansible code simpler.
- ansible/roles/accumulo/files/accumulo-cluster-systemd: this might need to be turned into a jinja template so the tracer sections can be included conditionally
Looking into doing this.
- ansible/roles/accumulo/tasks/main.yml: don't copy the tracers file for >= 2.1.0
- ansible/roles/accumulo/templates/accumulo-env.sh: remove tracer from the accumulo 2.1.0 case
Did these in 7dcc217
The build is failing. May be running into this issue https://github.com/ansible-community/ansible-lint/issues/1795
The build is failing. May be running into this issue ansible-community/ansible-lint#1795
Strange... the Python version is still 3.9.9, so that has not changed. Anyways looks like bumping the version of ansible-lint is a simple fix.
Strange... the Python version is still 3.9.9, so that has not changed. Anyways looks like bumping the version of ansible-lint is a simple fix.
Not completely sure, but I think the new version of ansible lint may have the following which fixes the issue with breaking change in one of its dependencies.
Strange... the Python version is still 3.9.9, so that has not changed. Anyways looks like bumping the version of ansible-lint is a simple fix.
Not completely sure, but I think the new version of ansible lint may have the following which fixes the issue with breaking change in one of its dependencies.
I'll double-check.
Trying to start the tracer for Accumulo 2.1.0-SNAPSHOT was causing problems. Changed code to avoid doing that.