8th of October, 2020
|
Download the latest version of Structure and its Extensions
Try It: Structure Sandbox Server (no installation required)
Version HighlightsStructure 6.3 adds:
Changes in DetailStatus Rollup EffectorThis new effector allows you to automatically update parent issue statuses based on the earliest status of their sub-issues. Documentation: Status Rollup Effector Default ViewsIt is now possible to select a default view for each structure right from the Views menu. When set, users who open the structure for the first time will see the default view. Additionally, if there is no recently-selected view available for the structure, it will open with the default view. Documentation: Views Menu Group by Project Category and ScriptRunner Single Issue PickerIt is now possible to group items by Project Category and the ScriptRunner Single Issue Picker field, using Group Automation. To Group by Project Category, select Automation | Group by Text Attribute, and then select Group by Project Category. If you have ScriptRunner installed and have defined Single Issue Pickers, they will appear under the Automation | Group by... menu.
Documentation: Group Generators, ScriptRunner Improved Information CenterThe improved Info window now includes additional links and information about Structure extensions. Additional Updates
Supported VersionsStructure 6.3 and all extensions support Jira versions 7.13 or later. This release is for customers using Jira Server or Data Center (Jira Core, Jira Software, or Jira Service Desk). Compatible plugins versions:
Cloud customers can learn more about our products on the “Cloud” tab of our marketplace listing. Installation and Upgrade
Installing StructureIf your Jira server does not have Structure yet, the installation is simple:
Upgrading StructureThe upgrade procedure from versions 3.0–6.2 is simple:
Upgrading from Structure 2.9–2.11
Starting with version 6.0, Structure is no longer able to access the old Structure 2.x database, but old backup files are still supported. Therefore, you'll need to back up your Structure 2.x data before upgrading. The recommended upgrade procedure is as follows:
Enterprise Deployment NotesStructure 6.3 introduces a few changes and improvements especially important for large-scale Jira Server and Jira Data Center instances. Reading Custom Fields from the IndexStarting with Structure 6.3, the attribute system reads the values of numeric, date, and date-time custom fields from the Lucene index instead of the database. Reading a single field for many issues from the index is usually much faster than loading a lot of issue objects one by one from the database. Structure generators have been using this technique for a long time, and now we're bringing it to the attribute subsystem. This improvement is especially important for aggregate values like sums, where Structure sometimes needs to load the values for many more issues than the user sees on the screen to calculate the sum. If you have big structures with thousands of issues and you calculate aggregates based on numeric, date and date-time custom fields on those structures, we advise that you verify the performance of those aggregate columns in a staging environment before upgrading. Cluster Node Tracking in Structure Gadget in ConfluenceThe Structure gadget, unlike most other Jira gadgets, is a live web application, and it uses multiple REST requests to load the data and keep it up-to-date. When a Structure gadget is inside a Confluence page, those REST requests have to be proxied by Confluence: the gadget calls Confluence, which then calls Jira and forwards its reply back to the gadget. For each forwarded request, Confluence authenticates the current user, creating a new, single-use user session. In a Jira Data Center instance, the load balancer can route these requests to different cluster nodes, so it is possible, for example, that the gadget loads the structure from one node, but the values shown in the grid come from a different node. Structure was not designed to operate in these conditions. This problem was hard to notice in earlier versions of Structure, but since Structure 6.0, completely unrelated values may be shown in the grid, giving the illusion that a different structure is displayed. In Structure 6.3, we've updated the gadget to keep track of the cluster node: it remembers the node from which it received its first reply, and if later a reply comes from a different node, the gadget will ignore it and retry the request, hoping that the load balancer will redirect it to the original node. If the gadget is unable to reach the original node after multiple attempts, it will stop updating the data and display an error message. If the original node is down, reloading the page will reinitialize the gadget and "bind" it to a different node. Also, the gadget will always switch to read-only mode if it's on a Confluence page and the Jira instance is a Data Center instance. This solution is only a temporary workaround. We are still researching better ways of integrating Structure into Confluence pages. More Reliable Export Under LoadWe have seen a few cases where, on a busy system, exporting a structure to an Excel file or printable HTML would fail due to a lock timeout because the exporting thread would not be able to lock and refresh the structure for 30 seconds. Users would then have to retry or, in the most extreme cases, schedule their exports during a quiet period. In Structure 6.3, we've made the export process more robust: it will now wait for at least 10 minutes to let the other threads finish and get access to the structure. Also, after 2 minutes of waiting time, the exporter will try to switch to a "fast path": if there's already a recent version of this structure in the cache, it will be exported without waiting for the lock (accompanied by a warning that the exported version may be outdated). We hope that these changes improve users' experience when exporting large, complex structures under heavy load. Testing on Staging EnvironmentApart from the changes and suggestions above, there are no particular special areas of interest for load testing and stress testing Structure 6.3. We advise running the same testing procedures as you've done for previous upgrades. |