Subject: Re: .AppleDouble tests on OpenBSD (long...)
From: Matthew Keller (kellermg@potsdam.edu)
Date: Mon Apr 23 2001 - 18:45:16 EDT
Good stuff here- Good luck with the buckets!
On 23 Apr 2001 17:20:33 -0500, Karen A Swanberg wrote:
>
> Ok, I'm doing my best to run a thorough set of tests to figure out when
> the .AppleDouble permissions tank on OpenBSD.
>
> First of all, I only ever get the errors on the server. From the client
> side, the usage is seamless, Joe MacUser would never know there was a
> problem.
>
> The client is a PowerMac G3, running OS 8.0. I also have access to the
> same computer model with 8.1 and 8.6, if that will help, and I also have
> blue G4's running 9.0.4. I choose this one because it was physically
> closest, and I could see the console on the BSD machine. However, the
> 9.0.4 machines have SSH, so I can log in and watch /var/log/messages from
> there as well if necessary.
>
> For testing purposes, I created two new test directories:
>
> /home2/sticky_no chmod 0777; chown geology;chgrp guest
> /home2/sticky_yes chmod 1777; chown geology;chgrp guest
>
> in order to see if the sticky bit had anything to do with this. These are
> brand new directories, so no legacy .AppleDouble or .Parent files.
>
> I added them to /usr/local/etc/AppleVolumes.default, and sighuped inetd.
>
> They're being accessed under the general test account, which in this case
> is geology/guest (I don't allow the actual guest account).
>
> With sticky_no, I log in from the client, create a new directory
> (sticky-no-2), and copy in a file from the client desktop. No errors. I
> then dismounted the directory, remounted it, and launched the file (it's a
> gif opening in Photoshop). No errors. I then threw the file away, and got
> these errors:
>
>
> Apr 23 16:30:20 epidote afpd[25250]: setdirowner: chown 16777216/-1
> .AppleDouble/.Parent: Operation not permitted
> Apr 23 16:30:20 epidote afpd[25250]: setdirowner: chown 16777216/-1
> .AppleDouble: Operation not permitted
> Apr 23 16:30:20 epidote afpd[25250]: setdirowner: chown 16777216/-1
> .: Operation not permitted
> Apr 23 16:30:20 epidote afpd[25250]: setdirowner: chown -1/0
> .AppleDouble/.Parent: Operation not permitted
> Apr 23 16:30:20 epidote afpd[25250]: setdirowner: chown -1/0
> .AppleDouble/.Parent: Operation not permitted
>
> I have not yet emptied the trash.
>
> epidote# pwd
> /home2/sticky_no
> ls -als
> 1 drwxrwxrwx 6 geology guest 512 Apr 23 16:30 .
> 1 drwxr-xr-x 8 root wheel 512 Apr 23 16:23 ..
> 1 drwxrwxrwx 3 geology guest 512 Apr 23 16:29 .AppleDesktop
> 1 drwxrwxrwx 2 geology guest 512 Apr 23 16:29 .AppleDouble
> 1 drwx--Srwx 4 geology guest 512 Apr 23 16:30 Network Trash Folder
> 1 drwxrwsrwx 3 geology guest 512 Apr 23 16:30 sticky-no-2
>
> (note, I did not set the perms on Network Trash Folder, netatalk did, and
> I don't know why the stickybit is on sticky-no-2. It shouldn't be there.
> Hm...Didn't notice this until way late, right before I sent this message.)
>
> I'll do the same process with /home2/sticky_yes now.
>
> Exact same sequence, same errors, with sticky_yes
>
> epidote# pwd
> /home2/sticky_ye
>
> 1 drwxrwxrwt 5 geology guest 512 Apr 23 16:56 .
> 1 drwxr-xr-x 8 root wheel 512 Apr 23 16:23 ..
> 1 drwxrwxrwx 2 geology guest 512 Apr 23 16:56 .AppleDouble
> 1 drwx--Srwx 3 geology guest 512 Apr 23 16:56 Network Trash Folder
> 1 drwxrwsrwx 3 geology guest 512 Apr 23 16:52 sticky-yes-2
>
> -------
>
> I'll see if I can get errors without involking the Trashcan.
>
> I'm totally cleaning out /home2/sticky_yes and /home2/sticky_no.
>
> I mounted /home2/sticky_no, and created a new subdirectory, sticky-no-2.
> No errors. I opened up simpletext and created a simple text file. As soon
> as I tried to save it to sticky-no-2 (even before I hit 'save' but after
> going into the mounted subdirectroy in the save dialog) I got these
> errors:
>
> Apr 23 16:48:01 epidote afpd[13868]: setdirowner: chown 16777216/-1
> .AppleDouble/.Parent: Operation not permitted
> Apr 23 16:48:01 epidote afpd[13868]: setdirowner: chown 16777216/-1
> .AppleDouble: Operation not permitted
> Apr 23 16:48:01 epidote afpd[13868]: setdirowner: chown
> 16777216/-1 .: Operation not permitted
> Apr 23 16:48:01 epidote afpd[13868]: setdirowner: chown -1/0
> .AppleDouble/.Parent: Operation not permitted
> Apr 23 16:48:01 epidote afpd[13868]: setdirowner: chown
> 16777216/-1 .AppleDouble/.Parent: Operation not permitted
> Apr 23 16:48:01 epidote afpd[13868]: setdirowner: chown 16777216/-1
> .AppleDouble: Operation not permitted
> Apr 23 16:48:01 epidote afpd[13868]: setdirowner: chown 16777216/-1
> .: Operation not permitted
> Apr 23 16:48:01 epidote afpd[13868]: setdirowner: chown -1/0
> .AppleDouble/.Parent: Operation not permitted
>
>
> The exact same thing happens with Stick_yes.
>
>
> ----
>
> So I left the test file, and unmounted sticky_yes. I remounted it, and got
> one instance of:
>
> Apr 23 16:54:11 epidote afpd[22793]: setdirowner: chown -1/0
> .AppleDouble/.Parent: Operation not permitted
>
> And to make absolutely sure, the errors are occurring exactly when I
> choose the netatalk-volume in the save-as dialog. I checked again.
>
> ---
>
> Well, I don't know how useful that is, and as my roof is leaking at home,
> I'm off to battle the buckets. However, I am more than willing to test
> this ad nauseum for whomever wants to direct my tests, as I'm just
> shooting in the dark here, albeit carefully and methodically. If someone
> would like me to run specific tests, with specific Mac OS's or
> whathaveyou, I will.
>
> Cheers!
>
> -Karen
>
> * *
> Karen Swanberg |
> Network Admin. | Error: Sector not found--
> Dept. of Geology/Geophysics |search behind couch? (Y/N)
> 206 Pillsbury Hall |
> 310 Pillsbury Ave. SE |
> University of Mn | Enter any 11-digit prime
> Minneapolis, MN 55455 | number to continue...
> (612) 624-6541 |
> * *
>
>
>
>
--Matthew Keller Enterprise System Analyst Computing & Technology Services Information Services Division State University of NY at Potsdam Potsdam, NY USA
This archive was generated by hypermail 2b28 : Sun Oct 14 2001 - 03:04:37 EDT