Reputation: 13
In our company we recently upgraded to artifactory-pro:7.38.10 from version 6. To cleanup old artifacts we are using lavatory which runs an aql-search to identify the artifacts to be removed by filtering them by the date. This worked without issues our previous installation based on artifactory 6. Now after the upgrade artifactory frequently crashes with an OutOfMemoryError and the instance seems to require either significantly more memory than before or there is a memory leak. After further investigation it turned out that when problem is caused by running the aql-search and the memory usage jumps from 4 GB to over 10 GB. That's +6 GB for something that hasn't changed.
After searching for known issue I found https://www.jfrog.com/jira/browse/RTFACT-26825 which is resolved and might solve our problem but there is no version specified containing a fix. Since there is a workaround and the issue was fixed, I expect that there must be a release.
Is there already a release containing a fix?
Upvotes: 0
Views: 157
Reputation: 579
The JIRA that you are referring to is fixed in the Artifactory version 7.38.0. So mostly, that should not be causing an issue as you are in higher version than 7.38.0.
In order to confirm, you may try the below. Add the system property to $JFROG_HOME/artifactory/var/etc/artifactory/artifactory.system.properties file and restart Artifactory for changes to take effect:
artifactory.nuget.v2.search.page.size=1000
Alternatively, you may put all the Nuget Devexpress repositories as offline. Now, check if you are encountering the memory issue. If you are not encountering this issue, maybe there is a regression issue. But according to my assumption, your server needs more resources as there are a lot of microservices introduced in Artifactory 7, when compared to Artifactory 6. Please check if you are satisfying the resources requirement as mentioned in this page. In that case, you would need to tune your Artifactory as per this article.
Upvotes: 0