Rundeck Community

Hi Team,I am setting up rundeck with Hashi Corp Vault, as part of setup process and testing found below issues, If anyone overcame these, please let me know the workaround steps.Rundeck version installed: 5.10.0Issues: I tired installing vault plugin, but rundeck did not recognize this plugin, so i ended up installing from Rundeck UI. Vault Storage Plugin version installed: 1.2As per rundeck documentation, this supports KV Engine V2, if we add parameter “rundeck.storage.provider.1.config.engineVersion=2” but this plugin is still using only KV engine V1. Once all the configuration is completed, restarted rundeck service, and was able to make successful connection to vault and able to see the keys, but as per our environment, we kept the vault token expire for 5 mins, once this expired rundeck is not auto using the new vault automatically. When rundeck lost connection to the vault due to token expiry, all jobs are failing irrespective if the jobs are using the keys from vault in the jobs

Previously I reported a CVE associated with Log4J Log4jLink.  This was resolved with the release of 6.0.  Thanks!  Almost immediately after I upgraded to 6.0, I got flagged on a new CVE:  Plugin Name:  Apache Log4j 2.13.1 < 2.25.5 / 2.26.0 < 2.26.1 Improper JSON Serialization (CVE-2026-49844)Plugin Output:  Path              : /var/lib/rundeck/bootstrap/rundeck-6.0.0-20260629.war  Installed version : 2.25.4  Fixed version     : 2.25.5Is there a plan for upgrading the Log4J version again?  Please advise. Thanks,Eric

Hello.I noticed there is a weird difference between the two timestamps shown in Rundeck 5.19 web interface.I’m referring tot he HH.mm:ss timestamp shown in the left sidebar, which I assume to be the server’s timestamp in local timezone and the ISO8601 timestamp shown in tooltip/.rdlog file. The problem is they look inconsistent so I was wondering how each of these timestamps is determined The timestamp form tooltip/.rdlog file has a Zulu suffix which suggest UTC but there is a huge inconsistencyThe server is located in France (UTC + 2), locale timezone is set correctly in the OS, so is the UTC clock. NTP synchronization is active. Therefore on Rundeck server, both local timezone and UTC are set correctly.Rundeck JVM is executed with the correct timezone parameter “-Duser.timezone=Europe/Paris”. But the problem is there is much more than 2 hours differences between the left sidebar timestamp and the ISO8601 tooltip timestamps.I noticed the issue issue in job during it’s execution on the

Within Rundeck, is it possible to restrict the ip source/locations for an api key to be used?  For instance, the “admin” userid’s api key can only be used by scripts/calls from the Rundeck server itself.  Any suggestions on how to restrict the source calling location of specific api keys (as normal regular end-users’ api keys can be from their laptops and such)?

After setting rundeck.security.apiCookieAccess.enabled to value of “false”; the “Projects” list pull-down shows an empty list after user’s log into the Rundeck gui.  Granted, the Rundeck home page still shows all the projects.  (Rundeck version 5.20.0)

After upgrading to Rundeck 5.20.0; it looks like the ssj-ssh plugin no longer works with double-quoted commands in the job steps.  For example, the following used to work:<exec>sudo "sh /u01/scripts/sre/myscript.sh ${option.parm1}"</exec> We had to change it to the following in order for it to work:<exec>sudo sh /u01/scripts/sre/myscript.sh ${option.parm1}</exec>With double-quotes, the job returns: [sudo] password for user1:sudo: sh /u01/scripts/sre/myscript.sh userval1: command not found[user1@serverxyz ~]$ echo $?1 Note: This is on Oracle Linux 9 server.

After upgrading to Rundeck 5.20.0 release; we see the following errors repeatedly in service.log:[2026-06-09T15:59:50,795] ERROR spi.SqlExceptionHelper - Unsupported conversion from DECIMAL to java.sql.Time[2026-06-09T15:59:50,975] WARN  spi.SqlExceptionHelper - SQL Error: 0, SQLState: 22018[2026-06-09T15:59:50,975] WARN  spi.SqlExceptionHelper - SQL Error: 0, SQLState: 22018 This is Rundeck 5.20.0 running on Oracle Linux 9.  The database being used is 8.0.46-commercial and Rundeck is connectiving via mysql-connector-j-9.7.0.jar .  Any suggestions?

We were recently doing a review of our plugins and noticed that version 2.0.0 is available for the Slack notification plugin. However when checking via “Find Plugins” the community version of Rundeck only shows version 1.2.4. Now from seeing the Rundeck documentation I see the Slack notification plugin is now a Enterprise plugin. So am I right in assuming the Community version will be stuck on 1.2.4?

Hello and welcome! Whether you are a long-time Rundeck user, a brand-new contributor, or just exploring what’s possible with runbook automation, we are thrilled to have you here.This dedicated section of the PagerDuty forum is a space for all things Rundeck. To keep this community helpful, healthy, and high-powered, we have a few simple guidelines:🤝 Code of Conduct (The Quick Version) Ask & Learn: No question is too basic. Ask about configurations, jobs, or specific use cases. Share the Knowledge: Discovered a cool workaround? Built a great workflow? Share your tips and tricks with the class! Be Helpful & Respectful: We are all here to learn from one another. Lift each other up and keep discussions constructive. 📚 Essential Rundeck ResourcesGot questions or want to dive into the code? Bookmark these essential links: Official Documentation: The ultimate guide for setup, configuration, and feature guides. Check out the Rundeck Docs. Found a Bug?: Help us make Rundeck bett

Hi,This is a WAR installation on FreeBSD, I added :# Allow donwload of public keysrundeck.gui.keystorage.downloadenabled=truerundeck.feature.publicKeysDownload.enabled=trueto both framework.properties and rundeck-config.properties, which I has to create (it doesn’t exist in FreeBSD) with no success.I also added this to preferences.properties Still doesn’t workRegards,Xavier

Hello,I introduced some times ago a workflow plugin, allowing to create dependency links between jobs that can subsist over the 24h duration of the workflow, which is defined with a daily basis.This can lead, for example, with a job failing for any reason, and the jobs linked to it waiting patiently until the error is fixed, and relaunched. If needed, waiting for the whole night.Since, I’ve worked on tackling the main problem due to linking dozen of jobs together : a visual representation of the jobs and the links between them, through a new UI plugin.Preferably, with an interactive output and able to be updated at will.I’ve reached a satisfactory level with it, where while the desired goal is reached, the UI plugin became more than useful even for people not using the other dependencies modules.To the point it is becoming a companion of the main UI.The github repo is available here, with the usage documentation.As the new module in the plugin is mostly on the browser side, the install

We recently upgraded to Rundeck 5.20 and set “rundeck.feature.nextUiMode.enabled = true”.  While the projects’ job lists load quicker (especially those with numerous cascading folders for job organization), the top of the screen shows duplicate search bars now.

Hello,Some of our users noticed a minor issue with "Job reference" step. There step label value is ignored. When a step label is added to a  "Job reference" step, the job can edition can be normally validated without any message error but the label value just disappears suddenly. It doesn't even appear after saving the stop (before saving the workflow)This happens in both Rundeck Pro/Enterprise and Rundeck Community…Is this a known bug?

Hello,When you add a "Job reference" step to a job, and choose “By name” to manually enter referenced the job’s group/name (rather then using “Choose a job” button and selecting a job interactively), is there supposed to be an autocomplete in jobs names/groups/names ? In both Rundeck Community and Pager Duty

We have enabled the execution history cleanup in Rundeck; however, the executions are not getting deleted as expected.As per the Rundeck documentation, it is mentioned that the cleanup feature needs to be enabled and properly configured. Could you please confirm:Is enabling and configuring the execution cleanup sufficient for automatic deletion? Or do we need to create a separate job to handle execution cleanup?We would like your guidance on the correct approach to ensure execution history is cleaned up as expected.And rundeck version is 3.3.11

Hello,Is there a way to reference a job from another WITHOUT the referenced job’s project name ?Some of our users use the SCM plugin to push job definition to a GitLab instance and share the referenced same jobs between two Rundeck instances used for two different purposes. Lets call them Instance A and Instance B.Instance A has project_X containing a referenced job_1, from job_2. Both jobs definitions are pushed to GitLab instance and cloned to project_Y on Instance BThe problem is, while the UUID and jobs name are the same on both Instance A and Instance B because both jobs do the same EXACT thing as you would except (no modification on actual jobs, they different only in terms of which nodes they’re scheduled/executed on), projects names on tho other hand reflect the purpose/environment (testing, production, level-1 support…).That naming scheme leads to different projects name even when the jobs are the similar in terms of jobs contents (steps, scripts, commands…), thus project_Y in

Hi,The table ‘stored_events’ grow exponentially, which each login of an user. Is it safe to purge this table ?Regards,Xavier

Hello,Our users reported many of their jobs fail to run when scheduled with the following error… com.dtolabs.rundeck.core.NodesetEmptyException: No matched nodes: NodeSet{includes={hostname=node_fqdn.tld, dominant=false, }} at com.dtolabs.rundeck.core.execution.workflow.BaseWorkflowExecutor.validateNodeSet(BaseWorkflowExecutor.java:880) at com.dtolabs.rundeck.core.execution.workflow.NodeFirstWorkflowExecutor.executeWorkflowImpl(NodeFirstWorkflowExecutor.java:92) at com.dtolabs.rundeck.core.execution.workflow.BaseWorkflowExecutor.executeWorkflow(BaseWorkflowExecutor.java:220) at com.dtolabs.rundeck.core.execution.WorkflowExecutionServiceThread.runWorkflow(WorkflowExecutionServiceThread.java:95) at com.dtolabs.rundeck.core.logging.LoggingManagerImpl$MyPluginLoggingManager.runWith(LoggingManagerImpl.java:146) at com.dtolabs.rundeck.core.execution.WorkflowExecutionServiceThread.run(WorkflowExecutionServiceThread.java:77)Exception: class com.dtolabs.rundeck.core.NodesetEmptyException: No ma

Hi all,On the previous forum, I posted:Hi All, Is there an upgrade available yet for the following vulnerability in the rundeck log4j library? Plugin Name: Apache Log4j 2.0-beta9 < 2.25.3 MitMPlugin ID: 282519Plugin Output:  Path              : /var/lib/rundeck/bootstrap/rundeck-5.12.0-20250512.war  Installed version : 2.17.2  Fixed version     : 2.25.3 I'm running rundeck 5.12.0. Thanks,Eric Racuna responded: Hi Eric, The engineering team is aware of that CVE, thanks for your feedback! Stay tuned to next releases. Regards! I checked in both 5.19.0 and 5.20.0 but it’s still not there.  Does anyone know when this is coming? jar tvf rundeck-5.20.0-20260402.war | grep log4j 24248 Mon May 12 16:40:32 CDT 2025 WEB-INF/lib/log4j-slf4j-impl-2.17.2.jar1811089 Mon May 12 16:40:32 CDT 2025 WEB-INF/lib/log4j-core-2.17.2.jar 30948 Mon May 12 16:40:32 CDT 2025 WEB-INF/lib/log4j-jul-2.17.2.jar302511 Mon May 12 16:40:34 CDT 2025 WEB-INF/lib/log4j-api-2.17.2.jar Thanks,Eric

Hi,Has the database structure changed between 5.19 and 5.20 ? We experience a regressions issue and want to try to rollback to 5.19Xavier

Hello.While looking at one of our users jobs, I noticed Rundeck Community 5.19.0 interface doesn’t show all the logs in node view. Here’s an example in screenshot Ansible playbook executed by a Rundeck job’s step from a distant node.Rundeck “Nodes” view is broken, with empty space instead of logsSometimes more logs is displayed (a few seconds more) but never the whole logs. The job actually ends a while after that.Note that the same issue as the missing “whale of a log” link for huge log files. This happens with very small log execution files. The corresponding .rdlog file is only 337 kB and both the .rdlog file and “Log output” view show all the logs but “Log output” view doesn’t lose the nodes  context (what step has been executed in which notes) and users don’t have access to Rundeck server’s filesystem and thus can’t access to .rdlog files of their jobs’ executions.Also, for some reason, all the steps (Ansible playbook, most of it being cloning git repos) are considered as errors.

Hello, is there a way to "Kill tracked processes after execution" the PID child when i kill a job that running for a specific command(Not a job)? 

Calling out all Rundeck by PagerDuty open-source experts out there!We’re gearing up for next Rundeck by PagerDuty meetups (in-person & virtual), and we want YOU to be one of our rockstar speakers! If you’ve got some cool insights, epic stories, or game-changing tips about your experience with Rundeck by PagerDuty open-source, we’d love to hear from you. This is your chance to shine, share your knowledge, and inspire your fellow community members. Plus, it’s a great way to network and connect with like-minded folks in the industry!If you’re interested (and we hope you are!), visit https://github.com/rundeck/rundeck-meetups, reach out to us at community-team@pagerduty.com, or leave a comment below and let us know your region, topic idea and availability. We can’t wait to make the next meetups the best one yet with your help!Watch a previous Rundeck by PagerDuty EMEA meetup with guest speakers Diego Infiesta from Ryanair and Hans Erasmus (@hans.erasmushe) from HBPS Consulting:  🎁 All