>BTW, it's really interesting to peek at the OpenAFS fileserver with  
>some of the dtrace tools.  It's *very* fdsync happy, and seems to  
>spend a lot of time/resources syncing buffers to disk -- I wonder if  
>we could seriously increase performance here and let the OS write- 
>cache actually be one...

Whew, you ain't kidding.  When I looked at that, I believe a lot of it
was the link table.  I have been idly thinking about simply removing
most of those fsync() calls, or collapsing a whole bunch of them ... it
would probably speed up operations like volume clones a whole lot.  A
few thought experiments made me think that perhaps the consequences of
an incorrect link count aren't so catastrophic that that salvager
couldn't easily recover from it ... but AFS has fooled me before, so
I'm not convinced of that yet :-)

--Ken
_______________________________________________
OpenAFS-devel mailing list
[email protected]
https://lists.openafs.org/mailman/listinfo/openafs-devel

Reply via email to