Problem Definition

Currently, Ranking operations are not possible on an issue if the issue has a Rank value with a length greater than 254 characters. Attempts to rank the issue will fail with errors such as the following:

2018-01-12 22:18:37,740 NodeReindexServiceThread:thread-1 WARN      [greenhopper.customfield.lexorank.LexoRankCFType] Unable to retrieve rank for field [15200] and issue [3867884]
2018-01-12 22:18:37,740 NodeReindexServiceThread:thread-1 WARN      [greenhopper.customfield.lexorank.LexoRankCFType] gh.api.rank.error.lexorank.fieldlength.exceeded.norebalance
2018-01-12 22:18:37,740 NodeReindexServiceThread:thread-1 WARN      [greenhopper.customfield.lexorank.LexoRankIndexer] Could not retrieve LexoRank value for issue[id=3867884]. Indexing max LexoRank value instead.

Issues will not be rankable again until a rebalance is complete, which can take a prohibitive amount of time depending on the size of the instance.

This is a static value configured in the source code of JIRA and cannot be adjusted in the UI, nor can it be configured through any configuration files or startup arguments.

As instances grow, the likelihood of encountering this rank length limit increases. In certain cases, and only under careful consideration from the administrators, it would be helpful to be allowed to adjust this value to a higher number to work around the limitation while the rebalance proceeds. Alternatively, it would be helpful to be allowed to bypass the limitation.

Suggested Solution

• Allow for administrators to be able to configure the maximum rank length via the UI, or at least via a configuration that can be set upon restarting JIRA.
• Or, allow administrators to bypass the rank limit in order to be able to continue ranking issues while the balance is ongoing.

Why this is important

Larger instances can face balance times of days, or even weeks. If this rank length limit is encountered, users will be unable to rank issues until the balance is complete. Administrators need a way to mitigate this.

Notes

It is worth noting that in instances that are frequently finding that their rank length is reaching the limit threshold, increasing the rank length would only be a workaround. To address the root cause and decrease the frequency when this occurs:

• JSWSERVER-16466 : Jira Agile getRankRelativeToOtherIssue() method allocate next Rank very inefficiently in some cases