Subject: Re: HELP: Request for Information!
From: Marcus Radich (marcus@darena.co.nz)
Date: Tue Sep 18 2001 - 05:34:22 EDT
Greetings.
My company supplies IBM x200 servers preconfigured to netatalk and
samba for Mac and PC publishing environments.
Our netatalk drama:
- We were using asun2.13 for a long time, but then with the intro of
the G4 - we had a problem with disconnects. I emailed adrian and he
told me to use 2.14 - he had just solved the disconnect issue with
2.14.
- We installed 2.14 on our problem servers, and it solved the
disconnect issue! Great......but.....then some operators were
complaining that their files were "already open by another user" etc
etc.
- Read through adrian's notes in the ReadMe and found this line:
" byte locks should now work. if you want to enable the old way
of doing things, add -DUSE_FLOCK_LOCKS."
I added this to our compile options - and the Quark 'open' problems went away.
This was all back in late 1999, and we have used asun2.14-37b since.
We have over 30 servers running with this config, and all works well.
We used to have the occasional problem with folders ending up in the
trash - but since switching to ReiserFS - no problems have shown up
(Fingers crossed).
Here is our latest config:
Hardware: IBM x200 e-series, Promise Ultra-100 ide card, IBM Disks.
OS: Mandrake 8.0 with all updates. Software-RAID with ReiserFS.
netatalk version: asun2.14 pre37b
compile options: DEFS= -DUSE_FLOCK_LOCKS
Macs:
OS 8.1 through to 9.2.1
Appleshare client 3.8.3 minimum
QuarkXPress 3.32r5 and 4.11
Quark files saved with previews on.
Any tests anyone wants me to run, let me know.
Marcus
At 1:27 AM -0700 18/9/2001, Steve Freitas wrote:
>Hi folks,
>
>I'm trying to do a non-programmer's job of tracking down the culprit behind
>the sometimes difficult relationship between Quark and Netatalk.
>
>We've got a complicated set of datapoints. Could those with non-working and
>working Quark installations supply the mailing list with a few points of
>information?
>
>I know a few of you have done this, but I'd like to generate a fresh report.
>Information is needed such as...
>
>1. Netatalk version (+ any patches), plus compile-time options if you
>compiled it (specifically locking-related options). If you didn't compile
>it, note the specific binary package used so the package generator can tell
>us about it.
>
>2. Server operating system, with kernel version and filesystem type. If you
>know of any strange issues with these, let us know. (E.g., I believe Linux
>2.2's locking code was flaky, don't know if that's fixed on 2.4. TcX cited
>this as a reason for not supporting row-level locking on MySQL, I believe.)
>
>3. Client operating system and version
>
>4. Quark version
>
>5. Quark save method (with/without previews, any other options)
>
>6. Whether it works, doesn't work, or partially works. If it doesn't work,
>what does it report?
>
>Hopefully we can build a little matrix and find a clue or three.
>
>Thanks!
>
>Steve
-- ========================================================"Always ask the question, never assume the answer." Marcus Radich 1999
========================================================
This archive was generated by hypermail 2b28 : Sun Oct 14 2001 - 03:04:52 EDT