HAT | 2 Jan 2012 15:01
Picon
Gravatar

Re: Latest patch breaks cnid_dbd

Hi, I also post to ML.

Rod Yager <rod <at> yager.net.au>, Mon, 2 Jan 2012 23:15:08 +1100
> Hi,
> 
> After applying the latest patch, cnid_dbd gives an error
> 
> Jan  2 22:45:34 moses cnid_dbd[14295]: error closing DB environment after recovery: Invalid argument
> 
> and running 
> 
> [root <at> moses .AppleDB]# /usr/local/bin/dbd -ref /data/Hippolyta/
> 
> yields 
> 
> 
> BDB3004 individual cache size too large: maximum is 10TB
> BDB0055 illegal flag specified to DB_ENV->backup
> Jan 02 22:53:34.368054 [14492] {dbif.c:504} (E:CNID): error closing DB environment after recovery:
Invalid argument
> 
> 
> Rolling back the patch gets things working again.

I can't reproduce it.
Which patch has started the problem?

filename should be compared using lower case.
    bf00805586b8c539e9341a17537543bd218cc5c1
    Mon, 2 Jan 2012 02:37:54 +0000
(Continue reading)

HAT | 3 Jan 2012 03:39
Picon
Gravatar

Re: Latest patch breaks cnid_dbd

Comparison of filename is completely unrelated to this problem.
Only the version of BDB is the cause.

From: Rod Yager <rod <at> yager.net.au>, Date: Tue, 3 Jan 2012 08:31:15 +1100
> After further investigation, it seems that the problem is associated with my having updated to
BerkeleyDB5.3 and was just triggered by this patch.
> 
> It works fine with BerkeleyDB5.2  (And reverting to DB5.2 was the only way I could cause the rebuild of a
working cnid database)
> 
> But I can successfully update to BerkeleyDB5.3 AFTER the patch is applied.
> 
> Rod
> 
> 
> 
> 
> On 03/01/2012, at 1:01 AM, HAT wrote:
> 
>> Hi, I also post to ML.
>> 
>> Rod Yager <rod <at> yager.net.au>, Mon, 2 Jan 2012 23:15:08 +1100
>>> Hi,
>>> 
>>> After applying the latest patch, cnid_dbd gives an error
>>> 
>>> Jan  2 22:45:34 moses cnid_dbd[14295]: error closing DB environment after recovery: Invalid argument
>>> 
>>> and running 
>>> 
(Continue reading)

Frank Lahm | 3 Jan 2012 15:18

Re: Latest patch breaks cnid_dbd

Fixed for 2.2.2 in ad759a0.

-f

2012/1/3 HAT <hat <at> fa2.so-net.ne.jp>:
> Comparison of filename is completely unrelated to this problem.
> Only the version of BDB is the cause.
>
> From: Rod Yager <rod <at> yager.net.au>, Date: Tue, 3 Jan 2012 08:31:15 +1100
>> After further investigation, it seems that the problem is associated with my having updated to
BerkeleyDB5.3 and was just triggered by this patch.
>>
>> It works fine with BerkeleyDB5.2  (And reverting to DB5.2 was the only way I could cause the rebuild of a
working cnid database)
>>
>> But I can successfully update to BerkeleyDB5.3 AFTER the patch is applied.
>>
>> Rod
>>
>>
>>
>>
>> On 03/01/2012, at 1:01 AM, HAT wrote:
>>
>>> Hi, I also post to ML.
>>>
>>> Rod Yager <rod <at> yager.net.au>, Mon, 2 Jan 2012 23:15:08 +1100
>>>> Hi,
>>>>
>>>> After applying the latest patch, cnid_dbd gives an error
(Continue reading)

HAT | 5 Jan 2012 14:54
Picon
Gravatar

more simple netatalk.conf

Q: Why is netatalk disliked? 
A: Configuration seems to be difficult!

I think that conf files of netatalk *3.0* should be simplified.
We should prepare it while netatalk *2.2.X* is current stable version.

At first, let's consider netatalk.conf.
In this file, only two parameters are important.

    CNID_METAD_RUN=
    AFPD_RUN=

Other parameters should be removed from netatalk.conf of netatalk 3.0.
In case of *BSD and systemd sytetem, the netatalk.conf itself
become unnecessary.

The following parameters are removed certainly because AppleTalk
stuff is abolished.

    ATALKD_RUN=
    PAPD_RUN=
    TIMELORD_RUN=
    A2BOOT_RUN=
    ATALK_BGROUND=
    ATALK_ZONE=

The following parameters can be removed because they can be set up
in afpd.conf.

    ATALK_NAME=
(Continue reading)

Frank Lahm | 6 Jan 2012 12:04

Re: more simple netatalk.conf

Hi,

2012/1/5 HAT <hat <at> fa2.so-net.ne.jp>:
> Q: Why is netatalk disliked?
> A: Configuration seems to be difficult!

;) Yeah, let's get rid of netatalk.conf.

> I think that conf files of netatalk *3.0* should be simplified.
> We should prepare it while netatalk *2.2.X* is current stable version.
>
> At first, let's consider netatalk.conf.
> In this file, only two parameters are important.
>
>    CNID_METAD_RUN=
>    AFPD_RUN=

These can be removed too. afpd and cnid_metad are just supposed to run. period.

> Other parameters should be removed from netatalk.conf of netatalk 3.0.
> In case of *BSD and systemd sytetem, the netatalk.conf itself
> become unnecessary.
>
> The following parameters are removed certainly because AppleTalk
> stuff is abolished.
>
>    ATALKD_RUN=
>    PAPD_RUN=
>    TIMELORD_RUN=
>    A2BOOT_RUN=
(Continue reading)

HAT | 6 Jan 2012 15:33
Picon
Gravatar

Re: more simple netatalk.conf

Hi,

From: Frank Lahm <franklahm <at> googlemail.com>, Fri, 6 Jan 2012 12:04:02 +0100
> ;) Yeah, let's get rid of netatalk.conf.

Gentoo's package "netatalk-2.2.1-r2" has already gotten rid of netatalk.conf.
http://packages.gentoo.org/package/net-fs/netatalk
I can't merge because it's peculiar.

>>    CNID_METAD_RUN=
>>    AFPD_RUN=
> 
> These can be removed too. afpd and cnid_metad are just supposed to run. period.

It's necessary to rewrite all initscripts.
Is it one initscript per daemon?

>> ATALK_UNIX_CHARSET=
>> I'm going to consider how to setup in afpd.conf.
> 
> Afair this is the same as -unixcodepage in afpd.conf, isn't it ?

They look alike.
But they are not completely the same.
In addition, "ATALK_UNIX_CHARSET='LOCALE'" may not operate correctly.

I'm going to investigate these carefully. 

>> CNID_CONFIG=
>> Should we prepare cnid_metad.conf?
(Continue reading)

HAT | 6 Jan 2012 16:40
Picon
Gravatar

more simple AppleVolumes.system

17 Jun 2009, Subject: Type/Creator mapping is legacy, Frank Lahm said,

>> search for default_type in etc/afpd/file.c
> I haven't looked at the code but instead did some testing and indeed:
> we're mapping on the fly. How cool! But better don't count the
> required CPU cycles. ;o)

No one is interested in AppleVolumes.system any longer.

# mv config/AppleVolumes.system doc/AppleVolumes.system.sample
# touch config/AppleVolumes.system

What about?

--

-- 
HAT

------------------------------------------------------------------------------
Ridiculously easy VDI. With Citrix VDI-in-a-Box, you don't need a complex
infrastructure or vast IT resources to deliver seamless, secure access to
virtual desktops. With this all-in-one solution, easily deploy virtual 
desktops for less than the cost of PCs and save 60% on VDI infrastructure 
costs. Try it free! http://p.sf.net/sfu/Citrix-VDIinabox
Frank Lahm | 6 Jan 2012 18:06

Re: more simple AppleVolumes.system

2012/1/6 HAT <hat <at> fa2.so-net.ne.jp>:
> 17 Jun 2009, Subject: Type/Creator mapping is legacy, Frank Lahm said,
>
>>> search for default_type in etc/afpd/file.c
>> I haven't looked at the code but instead did some testing and indeed:
>> we're mapping on the fly. How cool! But better don't count the
>> required CPU cycles. ;o)
>
> No one is interested in AppleVolumes.system any longer.

Probably.

> # mv config/AppleVolumes.system doc/AppleVolumes.system.sample
> # touch config/AppleVolumes.system
>
> What about?

What about
 # rm config/AppleVolumes.system
removing it from the code too of course ?

-f

------------------------------------------------------------------------------
Ridiculously easy VDI. With Citrix VDI-in-a-Box, you don't need a complex
infrastructure or vast IT resources to deliver seamless, secure access to
virtual desktops. With this all-in-one solution, easily deploy virtual 
desktops for less than the cost of PCs and save 60% on VDI infrastructure 
costs. Try it free! http://p.sf.net/sfu/Citrix-VDIinabox
(Continue reading)

HAT | 6 Jan 2012 18:11
Picon
Gravatar

Re: more simple AppleVolumes.system

>> # mv config/AppleVolumes.system doc/AppleVolumes.system.sample
>> # touch config/AppleVolumes.system
>>
>> What about?
> 
> What about
>  # rm config/AppleVolumes.system
> removing it from the code too of course ?

agree.

--

-- 
HAT

------------------------------------------------------------------------------
Ridiculously easy VDI. With Citrix VDI-in-a-Box, you don't need a complex
infrastructure or vast IT resources to deliver seamless, secure access to
virtual desktops. With this all-in-one solution, easily deploy virtual 
desktops for less than the cost of PCs and save 60% on VDI infrastructure 
costs. Try it free! http://p.sf.net/sfu/Citrix-VDIinabox
Frank Lahm | 7 Jan 2012 08:32

Re: more simple netatalk.conf

Hi,

2012/1/6 HAT <hat <at> fa2.so-net.ne.jp>:
> From: Frank Lahm <franklahm <at> googlemail.com>, Fri, 6 Jan 2012 12:04:02 +0100
>>>    CNID_METAD_RUN=
>>>    AFPD_RUN=
>>
>> These can be removed too. afpd and cnid_metad are just supposed to run. period.
>
> It's necessary to rewrite all initscripts.
> Is it one initscript per daemon?

No. One initscript that starts all necessary daemons (afpd and cnid_metad).

>>> ATALK_UNIX_CHARSET=
>>> I'm going to consider how to setup in afpd.conf.
>>
>> Afair this is the same as -unixcodepage in afpd.conf, isn't it ?
>
> They look alike.
> But they are not completely the same.
> In addition, "ATALK_UNIX_CHARSET='LOCALE'" may not operate correctly.
>
> I'm going to investigate these carefully.

Great!

>>> CNID_CONFIG=
>>> Should we prepare cnid_metad.conf?
>>
(Continue reading)


Gmane