Score:1

ReFS scheduled tasks?

sy flag

Does ReFS do any kind of automated tasks (reindex, consistency check, something like that) and if so, what/where/when?

I have a Server 2019 system that locks up every Saturday at midnight due to ReFS-related processes taking up something like half the physical memory via the non-paged memory pool (i.e. the pool usage increases massively and most of the high-usage pool tags are directly connected to refs.sys when I analyze the memory dump). The rest is taken up by the "metafile" (as seen by RAMMap) - I assume it's also ReFS related but haven't found a way to know for sure yet. Between them physical memory usage runs up to 99/100% and after a short time the OS locks up.

Since the issue only occurs once a week at the same time every week, it seems likely that some kind of scheduled weekly task is running. However, I cannot find any such task in Task Scheduler, and thus I wondered if it was an inbuilt (i.e. unrelated to Task Scheduler) ReFS task.

Thanks!

cn flag
You have integrity streams on? I think that is a once per week check that all copies / parts of the data are good. Then deduplication may actually use a lot of ram and runs also on a schedule, if deduplication is on.
djdomi avatar
za flag
does 2019 support dedup on Refs? however usually get-dedup job would be a good point and get-dedupschedule
DB_2022 avatar
sy flag
Thanks for the replies. It looks like Integrity Streams is on since the Data Integrity Scan runs every Saturday - but it runs in the morning, not at midnight. Although I don't know what it's scanning since the ReFS drive and DPM folders come up with Integrity Streams not enabled when I query with PowerShell. Same for deduplication - I don't see any volumes with dedup turned on but the relevant Scheduled Tasks run every day/week. But again, they don't run at midnight on Saturday.
I sit in a Tesla and translated this thread with Ai:

mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.