Frank Lahm | 5 May 10:00 2009

Re: default UAM

2009/4/30 Frank Lahm <franklahm <at> googlemail.com>:
> 2009/4/30 HAT <hat <at> fa2.so-net.ne.jp>:
>>>There are some issues with the patch:
>>>
>>>distrib/initscripts/rc.atalk.debian.tmpl:
>>>  We know that these guys can't build dhx, so we shouldn't put it into
>>>AFPD_UAMLIST
>>>  (although afair Debian doesn't use our intiscript). I'll adjust this mys
>>>elf.
>>
>> Debian .deb package cannot contain OpenSSL on account of the license.
>> But users can build with OpenSSL.
>> It is necessary to include DHX2 in tmpl file so that users may refer.
>>
>> Index: distrib/initscripts/rc.atalk.debian.tmpl
>> ===================================================================
>> RCS file: /cvsroot/netatalk/netatalk/distrib/initscripts/rc.atalk.debian.tmpl,v
>> retrieving revision 1.1.4.2
>> diff -u -r1.1.4.2 rc.atalk.debian.tmpl
>> --- distrib/initscripts/rc.atalk.debian.tmpl    30 Apr 2009 09:35:06 -0000      1.1.4.2
>> +++ distrib/initscripts/rc.atalk.debian.tmpl    30 Apr 2009 13:21:08 -0000
>>  <at>  <at>  -17,7 +17,7  <at>  <at> 
>>  test -x :SBINDIR:/atalkd || exit 0
>>
>>  # Set defaults. Please change these options in :ETCDIR:/netatalk.conf.
>> -AFPD_UAMLIST="-U uams_dhx.so"
>> +AFPD_UAMLIST="-U uams_dhx.so,uams_dhx2.so"
>>  AFPD_GUEST=nobody
>>  AFPD_MAX_CLIENTS=50
>>  ATALK_ZONE=
(Continue reading)

Frank Lahm | 5 May 10:21 2009

Re: doc/DEVELOPER

Hi,

2009/4/30 HAT <hat <at> fa2.so-net.ne.jp>:
> Hi,
>
> in doc/DEVELOPER,
>
> * should add "AFP over TCP".
> * BDB's URL is old.
> * BDB's version is old.

Grmpf. I just saw that you attached a patch which I didn't notice
before. I'll commit, slightly modified though:
+Currently, Netatalk supports BDB 4.1.25 and later.
EOF

-Frank

------------------------------------------------------------------------------
The NEW KODAK i700 Series Scanners deliver under ANY circumstances! Your
production scanning environment may not be a perfect world - but thanks to
Kodak, there's a perfect scanner to get the job done! With the NEW KODAK i700
Series Scanner you'll get full speed at 300 dpi even with all image 
processing features enabled. http://p.sf.net/sfu/kodak-com
Frank Lahm | 5 May 10:34 2009

Re: Netatalk 2.0.4 status

2009/4/30 Frank Lahm <franklahm <at> googlemail.com>:
> 2009/4/26 Andrew Morgan <morgan <at> orst.edu>:
>> On Fri, 24 Apr 2009, Frank Lahm wrote:
>>
>>> I've mailed a rc1 tarball to Andrew.
>>
>> And I have posted it to SF.net and released it.  Enjoy!
>
> I've mailed rc2 to Andrew.

...once more.

-Frank

------------------------------------------------------------------------------
The NEW KODAK i700 Series Scanners deliver under ANY circumstances! Your
production scanning environment may not be a perfect world - but thanks to
Kodak, there's a perfect scanner to get the job done! With the NEW KODAK i700
Series Scanner you'll get full speed at 300 dpi even with all image 
processing features enabled. http://p.sf.net/sfu/kodak-com
_______________________________________________
Netatalk-devel mailing list
Netatalk-devel <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/netatalk-devel
Andrew Morgan | 5 May 22:52 2009

Re: Netatalk 2.0.4 status

On Tue, 5 May 2009, Frank Lahm wrote:

> 2009/4/30 Frank Lahm <franklahm <at> googlemail.com>:
>> 2009/4/26 Andrew Morgan <morgan <at> orst.edu>:
>>> On Fri, 24 Apr 2009, Frank Lahm wrote:
>>>
>>>> I've mailed a rc1 tarball to Andrew.
>>>
>>> And I have posted it to SF.net and released it.  Enjoy!
>>
>> I've mailed rc2 to Andrew.
>
> ...once more.

Once again, posted to SF.net and released!

 	Andy
------------------------------------------------------------------------------
The NEW KODAK i700 Series Scanners deliver under ANY circumstances! Your
production scanning environment may not be a perfect world - but thanks to
Kodak, there's a perfect scanner to get the job done! With the NEW KODAK i700
Series Scanner you'll get full speed at 300 dpi even with all image 
processing features enabled. http://p.sf.net/sfu/kodak-com
_______________________________________________
Netatalk-devel mailing list
Netatalk-devel <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/netatalk-devel
(Continue reading)

Frank Lahm | 6 May 13:57 2009

Re: [ANN] New utility to maintain dbd databases: dbd

I did another commit as I discovered a serious bug which caused
running cnid_dbds to exit after dbd was run on their volume.

dbd -s and -r now are capable of:
   St Type Check (Status of implementation progress, type: file/dir)
   OK D    Make sure .AppleDouble dir exist, create if missing.
   OK F    Make sure ad file exists
   OK F/D  Delete orphaned ad-files, log dirs in ad-dir
   OK F/D  Check name encoding by roundtripping, log on error

dbd -d now also dumps the rootinfo key.

------------------------------------------------------------------------------
The NEW KODAK i700 Series Scanners deliver under ANY circumstances! Your
production scanning environment may not be a perfect world - but thanks to
Kodak, there's a perfect scanner to get the job done! With the NEW KODAK i700
Series Scanner you'll get full speed at 300 dpi even with all image 
processing features enabled. http://p.sf.net/sfu/kodak-com
Frank Lahm | 14 May 15:56 2009

Re: [ANN] New utility to maintain dbd databases: dbd

Hi,

I'm finished with pass 1 of the dbd scanning mumbo-jumbo:
---8<---
  1st pass: Scan volume
  --------------------

  St Type Check
  -- ---- -----
  OK F/D  Make sure ad file exists
  OK D    Make sure .AppleDouble dir exist, create if missing. Error
creating
          it is fatal as that shouldn't happen as root.
  OK F/D  Delete orphaned ad-files, log dirs in ad-dir
  OK F/D  Check name encoding by roundtripping, log on error
  OK F/D  try: read CNID from ad file (if cnid caching is on)
          try: fetch CNID from database
          -> on mismatch: use CNID from file, update database
(deleting both found CNIDs first)
          -> if no CNID in ad file: write CNID from database to ad
file
          -> if no CNID in database: add CNID from ad file to database
          -> on no CNID at all: create one and store in both places
  OK F/D  Add found CNID, DID, filename, dev/inode, stamp to rebuild database
---8<---

If you test: be carefull, backup first!

-Frank

(Continue reading)

Frank Lahm | 15 May 09:15 2009

Re: [Netatalk-admins] Automatically disconnecting users

Hi,

2009/5/15 didier <dgautheron <at> magic.fr>:
> Le jeudi 14 mai 2009 à 10:43 -0400, Ian Levesque a écrit :
>> I was wondering what happens if you remove a share from
>> "AppleVolumes.default" that users are presently connected to  (and
>> then "kill -HUP <parent afpd pid>").  Does afpd disconnect the users
>> connected to the now nonexistent share or do they stick around until
>> they disconnect on their own?
>
> With 2.0.x they stick around, with cvs HEAD there's an afpd option
> '-closevol' default as 2.0.x if set disconnect clients.

interesting, I didn't know that one too.
I see it was added here: http://tinyurl.com/ou93pl

But how do we notify the client that his volume handle is invalid now ?

And it's not in the manpage. I can fix that, but what to put in there
? Something like:
"Immediatly unmount volumes removed from AppleVolumes files on SIGHUP
sent to the afp master process" ?

-Frank

------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensing option that enables 
unlimited royalty-free distribution of the report engine 
for externally facing server and web deployment. 
(Continue reading)

didier | 15 May 22:55 2009
Picon

Re: [Netatalk-admins] Automatically disconnecting users

Hi,
Le vendredi 15 mai 2009 à 09:15 +0200, Frank Lahm a écrit :
> Hi,
> 
> 2009/5/15 didier <dgautheron <at> magic.fr>:
> > Le jeudi 14 mai 2009 à 10:43 -0400, Ian Levesque a écrit :
> >> I was wondering what happens if you remove a share from
> >> "AppleVolumes.default" that users are presently connected to  (and
> >> then "kill -HUP <parent afpd pid>").  Does afpd disconnect the users
> >> connected to the now nonexistent share or do they stick around until
> >> they disconnect on their own?
> >
> > With 2.0.x they stick around, with cvs HEAD there's an afpd option
> > '-closevol' default as 2.0.x if set disconnect clients.
> 
> interesting, I didn't know that one too.
> I see it was added here: http://tinyurl.com/ou93pl
> 
> But how do we notify the client that his volume handle is invalid now ?
Easy, we don't :) Afpd returns AFPERR_PARAM because now the volume id is
invalid and hopes for the best.

> 
> And it's not in the manpage. I can fix that, but what to put in there
> ? Something like:
> "Immediatly unmount volumes removed from AppleVolumes files on SIGHUP
> sent to the afp master process" ?
Something like that but we are reaching the limit of SIGHUP and
configuration file as administrative tool.

(Continue reading)

HAT | 16 May 17:35 2009
Picon

current CVS HEAD segfault on Fedora 10

cnid_dbd segfault on Fedora 10 when a client connects.
why?

* /var/log/messages

May 17 00:09:37 pfedora10 kernel: cnid_dbd[1903]: segfault at 3 ip 007993f6 sp bfaf4044 error 4 in libc-2.9.so[720000+16e000]
May 17 00:09:37 pfedora10 afpd[1902]: dbd_rpc: Error reading header from fd (db_dir /home/hat):
Connection reset by peer
May 17 00:09:37 pfedora10 afpd[1902]: transmit: something went wrong

* AppleVolumes.default
:DEFAULT: cnidscheme:dbd maccharset:MAC_JAPANESE options:upriv,usedots dperm:550 fperm:440

* afpd.conf
- -noddp -maccodepage MAC_JAPANESE

* netatalk.conf
AFPD_MAX_CLIENTS=20
ATALK_NAME=`echo ${HOSTNAME}|cut -d. -f1`
ATALK_MAC_CHARSET='MAC_JAPANESE'
ATALK_UNIX_CHARSET='LOCALE'
AFPD_GUEST=nobody
ATALKD_RUN=no
PAPD_RUN=NO
CNID_METAD_RUN=yes
AFPD_RUN=yes
TIMELORD_RUN=no
A2BOOT_RUN=no
ATALK_BGROUND=no
export ATALK_MAC_CHARSET
(Continue reading)

Frank Lahm | 16 May 19:09 2009

Re: current CVS HEAD segfault on Fedora 10

Hi,

2009/5/16 HAT <hat <at> fa2.so-net.ne.jp>:
> cnid_dbd segfault on Fedora 10 when a client connects.
> why?

Does it segfault on all volumes or only on that one ?

HEAD has this cool new logger, please set up cnid_meta to max_debug
logging by adjusting the netatalk startup script. Add
-l "default log_maxdebug /var/log/cnid.log"

The "" are necessary.

Maybe that gives me a clue.

A SBT from a core file would be nice too. Do you know how to ge one?

-Frank

------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensing option that enables 
unlimited royalty-free distribution of the report engine 
for externally facing server and web deployment. 
http://p.sf.net/sfu/businessobjects

Gmane