-
Notifications
You must be signed in to change notification settings - Fork 49
systemd service opendistro-performance-analyzer.service has the wrong binary path configured #235
Comments
Thanks for finding this issue @perkons ! If you have a PR ready, we're happy to review and accept it. Contributions welcome! |
After this fix the service still fails with the error: Also, I am not sure what is done with the package build as the postrm script removes everything that postinst copies. Please look here https://github.com/opendistro-for-elasticsearch/performance-analyzer/blob/master/packaging/rpm/postrm#L13 Please note that I also only changed rpm build, I did not touch deb. |
Remember that %postuninstall runs on package upgrade as well. See https://docs.fedoraproject.org/en-US/packaging-guidelines/Scriptlets/ for more information. Not sure if this is the issue or not. rpm verify fails as well:
|
@opoplawski the package build in this case is pretty incorrect, there should be no use of mv and cp, instead the spec file should be used with something like this:
I just upgraded to 1.12.0 and the issue is still there, the binary is not here /usr/share/elasticsearch/bin/performance-analyzer-agent-cli. And even if I copy it manually and set exacutable, I get:
Also I do not understand there reason behind this https://github.com/opendistro-for-elasticsearch/performance-analyzer/blob/master/packaging/rpm/postrm as it removes everything. |
Hello! I'm getting the same error: Failed at step EXEC spawning /usr/share/elasticsearch/bin/performance-analyzer-agent-cli: No such file or directory Is there any workaround? Thanks. |
@gferrette you could copy the file manually and set exec:
unfortunatelly there is another problem. You will get a new error:
|
I just did redeploy my lab cluster. That means fresh installed CentOS 8 and Opendistro 1.12.0 and Performance Analyzer is working just fine. This proves my conclusion on the package build. It is wrong as you cannot upgrade the cluster without breaking components/plugins. |
We have also the same error after upgrading to v.1.12:
|
Reinstalling the opendistro-performance-analyzer package fixed the issue for me. After reinstallation, performance-analyzer-agent-cli is moved back to /usr/share/elasticsearch/bin/ which fixed the issue. |
OS, OpenDistro and Java version used:
Used the documentation here https://opendistro.github.io/for-elasticsearch-docs/docs/install/rpm/ to install rmp packages from repo https://d3g5vo6xdbdb9a.cloudfront.net/yum/opendistroforelasticsearch-artifacts.repo
First this problem appeared in opendistroforelasticsearch-1.10.1 (opendistro/for-elasticsearch#94)
What happened:
After upgrading OpenDistro from previous to 1.10.1 or 1.11.0 Performance Analyzer systemd service Failed to start.
This is a rpm package build issue as it seems that the binary path has changed from
/usr/share/elasticsearch/bin/performance-analyzer-agent-cli
to/usr/share/elasticsearch/plugins/opendistro_performance_analyzer/bin/performance-analyzer-agent-cli
but the systemd service file has not been updated.
The file
/usr/lib/systemd/system/opendistro-performance-analyzer.service
looks like this after install:but
/usr/share/elasticsearch/bin/performance-analyzer-agent-cli
does not exist.It has been moved to
/usr/share/elasticsearch/plugins/opendistro_performance_analyzer/bin/performance-analyzer-agent-cli
also it is not executable.
What you expected to happen:
Opendistro for Elasticsearch Performance Analyzer starts.
The text was updated successfully, but these errors were encountered: