我们 Discourse 实例的数据库并非本地容器，而是使用了一个集中式数据库集群。最近，中央 PostgreSQL 已更新至 18.3 版本。我注意到现在的 Discourse 备份失败了：
[2026-05-19 03:37:37] [STARTED]
[2026-05-19 03:37:37] 'system' 已启动备份！
[2026-05-19 03:37:37] 将备份标记为运行中...
[2026-05-19 03:37:37] 确保 '/var/www/discourse/tmp/backups/default/2026-05-19-033737' 存在...
[2026-05-19 03:37:37] 确保 '/var/www/discourse/public/backups/default' 存在...
[2026-05-19 03:37:37] 更新元数据...
[2026-05-19 03:37:37] 转储数据库的公共架构...
[2026-05-19 03:37:37] pg_dump: 错误：因服务器版本不匹配而中止
[2026-05-19 03:37:37] pg_dump: 详情：服务器版本：18.3 (Ubuntu 18.3-1.pgdg24.04+1)；pg_dump 版本：15.15 (Debian 15.15-1.pgdg12+1)
[2026-05-19 03:37:37] 异常：pg_dump 失败
[2026-05-19 03:37:37] /var/www/discourse/lib/backup_restore/creator.rb:173:in 'BackupRestore::Creator#dump_public_schema'
/var/www/discourse/lib/backup_restore/creator.rb:36:in 'BackupRestore::Creator#run'
/var/www/discourse/lib/backup_restore.rb:13:in 'BackupRestore.backup!'
/var/www/discourse/app/jobs/regular/create_backup.rb:10:in 'Jobs::CreateBackup#execute'
有什么修复建议吗？