• Open Menu Close Menu
  • Apple
  • Shopping Bag
  • Apple
  • Mac
  • iPad
  • iPhone
  • Watch
  • TV
  • Music
  • Support
  • Search apple.com
  • Shopping Bag

Lists

Open Menu Close Menu
  • Terms and Conditions
  • Lists hosted on this site
  • Email the Postmaster
  • Tips for posting to public mailing lists
Re: Seeking cooperation from mds and Spotlight
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Seeking cooperation from mds and Spotlight


  • Subject: Re: Seeking cooperation from mds and Spotlight
  • From: Alexander Dimitrov <email@hidden>
  • Date: Thu, 19 Dec 2013 01:12:07 +0200

Hi Lund,
I had worked on improving support for mds on one filesystem, which was implemented by other people and I can share what was needed to be modified before mds started working on it:
1. We had a bug in the memory mapped files support (pageout/pagein) which resulted in file corruptions and mds was heavily using mmaped files.I tested this by writing a couple of simple programs playing with memory mapped files and also by using the fsx[1] tool.



mds only really works if mounted with diskutil? Currently, we mount ZFS
with regular mount(2) call, and not via diskutil and Disk Arbitration.
Would a HFS+ filesystem mounted with mount(2) also fail with mds? Should we
create a fs.bundle, with zfs.util -k to reply with a UUID for DA, and use
diskutil to mount?

The mount command does not make the new mount available as a new volume to OSX, AFAIK. Only diskutil / diskarb do that. 
2. I can confirm that. I noticed that if you don't use Disk Arbitration then Spotlight won't notice your volume every time, so I needed to use DA to mount our volumes. 

And mfs relies on volumes. In fact, mds also relies on fsevents, in order to learn of file changes. It might well be that mount does not invoke fsevents to make mds aware of the new node. Just guessing here, again, though.
Yes, AFAIK mds relies on fsevents but they are posted at the end of the relevant syscalls. So the local filesystems get this feature for free. It is different game for network filesystems, though.

3. I speculated that if I mark the mount with MNT_DOVOLS flag then mds will expect the fs to provide persistent object IDs between mounts, but our fs didn't have such feature, so after disabling this flag mds liked us.

Hope this helps. Good luck!
Alex
 
[1] http://fstools.macosforge.org/trac/browser/trunk/src/fsx 

 _______________________________________________
Do not post admin requests to the list. They will be ignored.
Filesystem-dev mailing list      (email@hidden)
Help/Unsubscribe/Update your Subscription:

This email sent to email@hidden

  • Follow-Ups:
    • Re: Seeking cooperation from mds and Spotlight
      • From: Jorgen Lundman <email@hidden>
References: 
 >Seeking cooperation from mds and Spotlight (From: Jorgen Lundman <email@hidden>)
 >Re: Seeking cooperation from mds and Spotlight (From: Thomas Tempelmann <email@hidden>)

  • Prev by Date: Re: Seeking cooperation from mds and Spotlight
  • Next by Date: Re: Seeking cooperation from mds and Spotlight
  • Previous by thread: Re: Seeking cooperation from mds and Spotlight
  • Next by thread: Re: Seeking cooperation from mds and Spotlight
  • Index(es):
    • Date
    • Thread