Send Netdot-users mailing list submissions to
[email protected]
To subscribe or unsubscribe via the World Wide Web, visit
https://osl.uoregon.edu/mailman/listinfo/netdot-users
or, via email, send a message with subject or body 'help' to
[email protected]
You can reach the person managing the list at
[email protected]
When replying, please edit your Subject line so it is more specific
than "Re: Contents of Netdot-users digest..."
Today's Topics:
1. National characters (Roger Andersson)
2. Re: Seeking feedback about DBMS support (William Bulley)
3. Re: Seeking feedback about DBMS support (Phil Regnauld)
4. Re: Seeking feedback about DBMS support (Ebo Thompson)
5. Re: Seeking feedback about DBMS support (William Bulley)
6. Re: [Netdot-devel] Seeking feedback about DBMS support
(Carlos Vicente)
----------------------------------------------------------------------
Message: 1
Date: Tue, 25 Sep 2012 14:46:21 +0200
From: Roger Andersson <[email protected]>
Subject: [Netdot-users] National characters
To: [email protected]
Message-ID:
<ofe1aef264.d93d712e-onc1257a84.00438ca5-c1257a84.00463...@csc.com>
Content-Type: text/plain; charset="us-ascii"
I currently can't use non english charcters in Netdot. Some background
facts:
- I downloaded the 0.9.10 virtual server, just to get a jumpstart
regarding all the dependencies
- downloaded Netdot V1.01 package
- droped the old and created a new database through make
- installed V1.0.1
- created some devices through snmp discovery
- found out that I couldn't use national characters in site names,
comments and stuff like that. Netdot GUI just truncates the strings from
the first national character
- edited /etc/my.cnf and restarted mysql: Nothing changed
- installed the swedish language pack and rebuilt
- edited /etc/default/locale
- the browser I'm using announces itself as using swedish
I still have the same problem. Now I have ran out of ideas. Could someone
show a blind man the right way to go?
/Roger
Additional info:
mysql> status
--------------
mysql Ver 14.14 Distrib 5.1.54, for debian-linux-gnu (x86_64) using
readline 6.2
Connection id: 44
Current database: netdot
Current user: root@localhost
SSL: Not in use
Current pager: stdout
Using outfile: ''
Using delimiter: ;
Server version: 5.1.54-1ubuntu4 (Ubuntu)
Protocol version: 10
Connection: Localhost via UNIX socket
Server characterset: utf8
Db characterset: utf8
Client characterset: utf8
Conn. characterset: utf8
netdot@netdot01:/var/log$ locale -a
C
en_AG
en_AG.utf8
en_AU.utf8
en_BW.utf8
en_CA.utf8
en_DK.utf8
en_GB.utf8
en_HK.utf8
en_IE.utf8
en_IN
en_IN.utf8
en_NG
en_NG.utf8
en_NZ.utf8
en_PH.utf8
en_SG.utf8
en_US.utf8
en_ZA.utf8
en_ZW.utf8
POSIX
sv_FI.utf8
sv_SE.utf8
netdot@netdot01:~$ cat /etc/default/locale
LANG="sv_SE.UTF-8"
netdot@netdot01:~$
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://osl.uoregon.edu/pipermail/netdot-users/attachments/20120925/8d0782d4/attachment-0001.html
------------------------------
Message: 2
Date: Tue, 25 Sep 2012 08:52:49 -0400
From: William Bulley <[email protected]>
Subject: Re: [Netdot-users] Seeking feedback about DBMS support
To: Carlos Vicente <[email protected]>
Cc: netdot-users <[email protected]>, netdot-devel
<[email protected]>
Message-ID: <[email protected]>
Content-Type: text/plain; charset=us-ascii
According to Carlos Vicente <[email protected]> on Tue, 09/25/12 at 08:35:
>
> Care to comment why?
Sure. But these are my opinions only, not those of my organization.
I have only used MySQL myself - zero experience with Pg - don't want
to climb the learning curve (unless there are good reasons to change).
Compare with source code management. I started using RCS in the 1980's
and still use it today. That said, when OU changed to Git, I immediately
embraced Git, even though the Git "look and feel" differs from CVS, SVN,
etc. It turns out the learning curve was not too bad, and Git rules!
But the main reason, here at the UofM, is because our IT folks want to
support a small set of database engines. Today, as I understand is, they
support Oracle _only_ but they reluctantly agreed to support MySQL for
some applications. I doubt I could get them to accept Pg in addition
to their organization-wide support mandate for Oracle and MySQL.
I need them to support MySQL when we deploy NETDOT in production. The
likelihood of an approved one-off PostgreSQL (on at minimum two servers)
installation is going to be small, IMHO. :-)
Regards,
web...
--
William Bulley Email: [email protected]
72 characters width template ----------------------------------------->|
------------------------------
Message: 3
Date: Tue, 25 Sep 2012 14:55:30 +0200
From: Phil Regnauld <[email protected]>
Subject: Re: [Netdot-users] Seeking feedback about DBMS support
To: Vincent Magnin <[email protected]>
Cc: [email protected]
Message-ID: <[email protected]>
Content-Type: text/plain; charset=us-ascii
Vincent Magnin (Vincent.Magnin) writes:
> > My hunch is that most people don't care about the backend, and if they
> > do, they prefer Pg. Am I right?
>
> At this time, I use mysql (most likely because Netdot used it as
> default). But, if you have a *faster* backend, I would prefer it.
At first, there is probably no direct performance gain to be had by
switching backends. But having used PostgreSQL for large data storage/
retrieval projects over the past 10+ years, I can say the following:
- Parts of the Netdot code that needs to deal with certain data types
can be simplified because certain transformations are no longer
necessary, if the DB backend can handle that type natively ;
- Not having to ping pong the data back and forth between the main
application (Netdot Perl code) and the DB saves some time on
processing when having to deal with many objects ;
- Certain features in PgSQL make it much easier to implement very
large data sets using so-called "partitioning" without having to
start messing around with multiple tables and keeping track of
those somewhere.
Since this is the netdot-users (and not -devel) list, you may think:
"Why do I care ?". Well, to start with:
- Netdot CAN be faster on some operations, like cleaning up (deleting)
the forwarding tables / MAC addresses: something which takes a very long
time with MySQL on large installations, can be done in a matter of
seconds
- Less time is spent on maintaining or optimizing code that is duplicating
functionality that is better handled by the storage engine (PgSQL);
for instance, storing/searching/sorting v4/v6 data types.
- This frees up more time from the core development team (currently less
than 2.5 people :)
... Which translates to less bugs, more time spent developing cool
features, maybe a faster release cycle, and happier users :)
Side note: the above are probably possible in MySQL, but then it becomes
a matter of figuring out which backend, which version of MySQL, etc...
Plus, there is in-house expertise to help with the PostgreSQL transition,
including making it as painless as possible to migrate an existing
installation to a new backend.
------------------------------
Message: 4
Date: Tue, 25 Sep 2012 13:21:04 +0000
From: Ebo Thompson <[email protected]>
Subject: Re: [Netdot-users] Seeking feedback about DBMS support
To: Carlos Vicente <[email protected]>
Cc: netdot-users <[email protected]>, netdot-devel
<[email protected]>
Message-ID:
<CA+ojiv9RBUNHRgy+mHT18_K4a+iZ_Bjemyf=6rnsy2k4a9o...@mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"
Hi,
Postgres db. The overiding factor diafavouring mysql is simply oracle!?
On Sep 24, 2012 8:10 PM, "Carlos Vicente" <[email protected]> wrote:
> Dear Netdot community,
>
> One of our initial design goals for Netdot was to make it DB-agnostic.
> We aimed at supporting both MySQL and Postgres equally. However, our
> production systems at the University of Oregon have always been based on
> MySQL, so Pg support was somewhat lagging. It still is.
>
> The sad truth is that it takes considerable time and effort to support
> two backends.
>
> At a recent discussion we talked about the possibility of supporting
> only PostgreSQL in the future. There are a few things that make Pg
> attractive: native support for IPv4 and IPv6 addresses, table
> partitions, etc. Pg is also not owned by Oracle ;-)
>
> A decision like this can't be made lightly. We certainly do not want to
> alienate administrators out there.
>
> My hunch is that most people don't care about the backend, and if they
> do, they prefer Pg. Am I right?
>
> What do you all think?
>
> --
> cv
> _______________________________________________
> Netdot-users mailing list
> [email protected]
> https://osl.uoregon.edu/mailman/listinfo/netdot-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://osl.uoregon.edu/pipermail/netdot-users/attachments/20120925/374da304/attachment-0001.html
------------------------------
Message: 5
Date: Tue, 25 Sep 2012 09:22:33 -0400
From: William Bulley <[email protected]>
Subject: Re: [Netdot-users] Seeking feedback about DBMS support
To: Phil Regnauld <[email protected]>
Cc: [email protected], [email protected]
Message-ID: <[email protected]>
Content-Type: text/plain; charset=us-ascii
According to Phil Regnauld <[email protected]> on Tue, 09/25/12 at 08:55:
>
> At first, there is probably no direct performance gain to be had by
> switching backends. But having used PostgreSQL for large data storage/
> retrieval projects over the past 10+ years, I can say the following:
>
> - Parts of the Netdot code that needs to deal with certain data types
> can be simplified because certain transformations are no longer
> necessary, if the DB backend can handle that type natively ;
>
> - Not having to ping pong the data back and forth between the main
> application (Netdot Perl code) and the DB saves some time on
> processing when having to deal with many objects ;
>
> - Certain features in PgSQL make it much easier to implement very
> large data sets using so-called "partitioning" without having to
> start messing around with multiple tables and keeping track of
> those somewhere.
>
> Since this is the netdot-users (and not -devel) list, you may think:
> "Why do I care ?". Well, to start with:
>
> - Netdot CAN be faster on some operations, like cleaning up (deleting)
> the forwarding tables / MAC addresses: something which takes a very long
> time with MySQL on large installations, can be done in a matter of
> seconds
>
> - Less time is spent on maintaining or optimizing code that is duplicating
> functionality that is better handled by the storage engine (PgSQL);
> for instance, storing/searching/sorting v4/v6 data types.
>
> - This frees up more time from the core development team (currently less
> than 2.5 people :)
>
> ... Which translates to less bugs, more time spent developing cool
> features, maybe a faster release cycle, and happier users :)
>
> Side note: the above are probably possible in MySQL, but then it becomes
> a matter of figuring out which backend, which version of MySQL, etc...
>
> Plus, there is in-house expertise to help with the PostgreSQL transition,
> including making it as painless as possible to migrate an existing
> installation to a new backend.
It appears OU is interested in simplifying or streamlining their coding
and support efforts. OTOH it might be worth considering building upon a
different, more modern support library. Long term gain for short term pain.
Rather than debate the merits of MySQL and PostgreSQL, would it be worth
talking about replacing Class::DBI with a more modern and thoroughly
object oriented solution such as Rose::DB?
Rose::DB supports PostgreSQL, MySQL, SQLite, Informix, and Oracle. All
the variations in database types would vanish should NETDOT embrace the
Rose::DB suite of Perl modules. This is not likely to be accepted, but
I feel compelled to make the suggestion given that Rose::DB is high on
the recommended list of DBI-like solutions for modern Perl applications.
Regards,
web...
--
William Bulley Email: [email protected]
72 characters width template ----------------------------------------->|
------------------------------
Message: 6
Date: Tue, 25 Sep 2012 09:31:59 -0400
From: Carlos Vicente <[email protected]>
Subject: Re: [Netdot-users] [Netdot-devel] Seeking feedback about DBMS
support
To: William Bulley <[email protected]>
Cc: [email protected], [email protected]
Message-ID: <[email protected]>
Content-Type: text/plain; charset=ISO-8859-1
Hi Web,
On 9/25/12 9:22 AM, William Bulley wrote:
> Rather than debate the merits of MySQL and PostgreSQL, would it be worth
> talking about replacing Class::DBI with a more modern and thoroughly
> object oriented solution such as Rose::DB?
This is not a question of either/or. Yes, Class::DBI is outdated and
there are better Perl ORMs out there. But,
> All
> the variations in database types would vanish should NETDOT embrace > the
> Rose::DB suite of Perl modules
that's not true. I believe that Phil Regnauld explained it pretty well
in his recent message. It doesn't matter which ORM you use, you still
have the problems of native data representation, management of large
tables, schema upgrades, etc.
And another fact is that changing the backend is far easier than
changing the ORM at this point.
--
cv
------------------------------
_______________________________________________
Netdot-users mailing list
[email protected]
https://osl.uoregon.edu/mailman/listinfo/netdot-users
End of Netdot-users Digest, Vol 46, Issue 19
********************************************
