Uploaded image for project: 'Jira Data Center'
  1. Jira Data Center
  2. JRASERVER-73023

Server Error while assigning Issue or performing mentioned Search in comment

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Fixed
    • Icon: Highest Highest
    • 8.22.0
    • 8.19.1, 8.20.0, 8.20.1, 8.21.0, 8.20.3, 8.20.11
    • Issue - Fields

      Issue Summary

      Server Error while assigning Issue or performing mentioned Search in comment

      Steps to Reproduce

      1. Configure Jira 8.20.1 instance connected to mssql
      2. Enable nested groups for the local Jira directory
      3. Create a hierarchy of nested groups that were members of each other like the following:
        1. Nested16 -> nested15 -> nested14 and so forth
      4. Added several users as members to nested group 16.
      5. Configured project permissions to only allow assignable users from group nested1.
      6. Attempt to assign user or perform mention search in comment

      Expected Results

      User should be able to Assign user and also perform @mention search in Comment.

      Actual Results

      Assigning and performing mention search in comment fails with below error

      Error faced by user

      The below exception is thrown in the atlassian-jira.log file:

      Caused by: com.microsoft.sqlserver.jdbc.SQLServerException: The incoming request has too many parameters. The server supports a maximum of 2100 parameters. Reduce the number of parameters and resend the request. 

      Workaround 

      • Modify project permissions if it's just a few projects
      • Disable nested groups entirely if it's happening all over the place

      Additional Information 

      With the above environment in place we can see the following query is executed whenever the search occurs:

      (@P0 nvarchar(4000),@P1 nvarchar(4000),@P2 nvarchar(4000),@P3 nvarchar(4000),@P4 nvarchar(4000),@P5 nvarchar(4000),@P6 nvarchar(4000),@P7 nvarchar(4000),@P8 nvarchar(4000),@P9 nvarchar(4000),@P10 nvarchar(4000),@P11 nvarchar(4000),@P12 nvarchar(4000),@P13 nvarchar(4000),@P14 nvarchar(4000),@P15 nvarchar(4000),@P16 nvarchar(4000),@P17 bigint,@P18 int,@P19 int,@P20 bigint)select CWD_USER.id, CWD_USER.directory_id, CWD_USER.user_name, CWD_USER.lower_user_name, CWD_USER.active, CWD_USER.created_date, CWD_USER.updated_date, CWD_USER.first_name, CWD_USER.lower_first_name, CWD_USER.last_name, CWD_USER.lower_last_name, CWD_USER.display_name, CWD_USER.lower_display_name, CWD_USER.email_address, CWD_USER.lower_email_address, CWD_USER.credential, CWD_USER.deleted_externally, CWD_USER.external_id
      from dbo.cwd_user CWD_USER
      where (CWD_USER.lower_user_name in (select MEMBERSHIP.lower_child_name
      from dbo.cwd_membership MEMBERSHIP
      where MEMBERSHIP.membership_type = @P0 and (MEMBERSHIP.lower_parent_name in (@P1, @P2, @P3, @P4, @P5, @P6, @P7, @P8, @P9, @P10, @P11, @P12, @P13, @P14, @P15, @P16)))) and CWD_USER.directory_id = @P17 and CWD_USER.active = @P18
      order by CWD_USER.lower_display_name asc
      offset @P19 rows fetch next @P20 rows only
      

      The following portion of the where clause keeps expanding in number of parameters as the nested tree grows:

      and (MEMBERSHIP.lower_parent_name in (@P1, @P2, @P3, @P4, @P5, @P6, @P7, @P8, @P9, @P10, @P11, @P12, @P13, @P14, @P15, @P16)
      

      Here we've added in nested17 into the hierarchy:

      and (MEMBERSHIP.lower_parent_name in (@P1, @P2, @P3, @P4, @P5, @P6, @P7, @P8, @P9, @P10, @P11, @P12, @P13, @P14, @P15, @P16, @P17)
      

      And now we have 22 in the hierarchy:

      and (MEMBERSHIP.lower_parent_name in (@P1, @P2, @P3, @P4, @P5, @P6, @P7, @P8, @P9, @P10, @P11, @P12, @P13, @P14, @P15, @P16, @P17, @P18, @P19, @P20, @P21, @P22)
      

      We can continue to grow this list of parameters by adding new groups into the hierarchy until we hit:

      Caused by: com.microsoft.sqlserver.jdbc.SQLServerException: The incoming request has too many parameters. The server supports a maximum of 2100 parameters. Reduce the number of parameters and resend the request.
      

      Looks like we made a lot of changes to how the lookup occurs as a part of https://jira.atlassian.com/browse/JRASERVER-72101, which may have introduced this regression in cases of deep nested group relationships.

              mswinarski Maciej Swinarski (Inactive)
              1ece1773342d Sandip Shrivastava
              Votes:
              0 Vote for this issue
              Watchers:
              12 Start watching this issue

                Created:
                Updated:
                Resolved: