Sorry, the configuration no longer exists。 You can modify spark's java.io.tmpdir information to avoid KE temporary files being generated in the tmp directory.
On Fri, Jan 5, 2024 at 3:23 PM MINGMING GE <7mmi...@gmail.com> wrote: > This directory cannot be deleted manually。 > This folder will be cleared by the crontab job of the system's temporary > files. In centos, it is the /etc/cron.daily/tmpwatch scheduled task, which > retains temporary files for up to 10 days. > > You can change its directory kylin.query.schema.tmp-directory through the > following configuration. > > On Fri, Jan 5, 2024 at 3:06 PM Li, Can <c...@ebay.com.invalid> wrote: > >> 在load hive表时，spark会生成一些临时文件以及目录 >> 用于持久化表的metadata信息，这一部分在service里是有缓存的，我们在清理这些临时文件的时候发现，如果这些文件被删除了会导致load >> table失败， >> >> 但是如果重启服务就正常，想请问一下这一块生成的临时文件能不能删除，或者多久能将其清理掉，是否有自动清理的功能。如果不能手动删除，是否能更改其生成文件时的目录，该怎么配置 >> >
