List:General Discussion« Previous MessageNext Message »
From:Gelu Date:May 2 2002 5:16pm
Subject:Re: mysql speed concerns
View as plain text  
Hi Anders,

>How ever, this is probaly not
> an interesstign discussion for the rest of the people who subscribes to
> the mysql list.
I AGREE.

But... i consider this "place", offer a good opportunity to exchange
ideas,knowledge and experiences about how we using or create MySQL
application .
In this way we can learn reciprocal: some from the others.Of course, maybe
some answers can be considered wrong because it's possible to be applicable
in other circumstances(unfortunately for the user/programmer who have
problems) and not in what it's described.Are several reason for what is
happened.But one of this is CONFUSION.
To avoid any confusion or doubt , I feel is my duty to be more specific ,
for sake of  the people who are subscribed in the mysql list:
For this reason :
IPC  - InterProcess Communication mechanism in the Unix environment -
described by the W. Richard Stevens(and not only) at pg.482  in book
"Advanced programming in the Unix environment".PIPE,MESSAGE
QUEUES,SEMAPHORES AND SHARED MEMORY are CLASSICAL FORMS of IPC. STREAM
PIPE,NAMED STREAM PIPE are ADVANCED FORMS of IPC. SOCKET is a type of file
for network(or nonnetwork) communication between processes.
ipcs - a little Unix program and provide information about IPC facilities.
TCP/IP - it was,it is and always remain :Transmission Control Protocol (TCP)
and is intended for use as a highly
reliable host-to-host protocol between hosts in packet-switched computer
communication networks.RFC 793 prepared for DARPA in 1981.
So, it's a big difference between DATA COMMUNICATION and PROCESS
COMMUNICATION.
All this are not invented by me.

I apologize for any inconvenience

Sincerely,
Gelu Gogancea



_____________________________________________________
G.NET SOFTWARE COMPANY

Permanent e-mail address : gg@stripped
                                          gelugogancea@stripped
----- Original Message -----
From: Svensson, B.A.T. (HKG) <B.A.T.Svensson@stripped>
To: Gelu <ggelu@stripped>
Sent: Thursday, May 02, 2002 2:04 PM
Subject: RE: mysql speed concerns


> As well as semaphores, shared memory and messages queues not necessarily
> involves IPC - on the other hand the protocoll family TCP/IP is always
IPC.
>
> We are talking about two different issues, and maybe you missunderstodd
> what I did talk about in the first place. How ever, this is probaly not
> an interesstign discussion for the rest of the people who subscribes to
> the mysql list.
>
> Regards,
>
> file://Anders
>
>
> > -----Original Message-----
> > From: Gelu [mailto:ggelu@stripped]
> > Sent: Thursday, May 02, 2002 9:52 AM
> > To: Michael Chang
> > Cc: Svensson, B.A.T. (HKG); Jason Yates; mysql@stripped
> > Subject: Re: mysql speed concerns
> >
> >
> > TCP/IP is a protocol for data interchange, host to host, on the network
> > (RFC731).
>
> > Sockets ,indeed , are also used for IPC.
> > _____________________________________________________
> > G.NET SOFTWARE COMPANY
> >
> > Permanent e-mail address : gg@stripped
> >                                           gelugogancea@stripped
> >
> > ----- Original Message -----
> > From: Michael Chang <miranda@stripped>
> > To: Gelu <ggelu@stripped>
> > Cc: Svensson, B.A.T. (HKG) <B.A.T.Svensson@stripped>; Jason Yates
> > <jyates@stripped>; <mysql@stripped>
> > Sent: Wednesday, May 01, 2002 10:29 PM
> > Subject: Re: mysql speed concerns
> >
> >
> > >
> > > Technically speaking, TCP/IP communication *is* a form of IPC.
> > > IPC isn't restricted to semaphores or message queues, etc.
> > > I think that's what he meant (i.e.: mysql client on one box
communicating
> > > via TCP/IP to a MySQL server on another box).
> > >
> > >
> > > Michael
> > >
> > >
> > >
> > > On Wed, 1 May 2002, Gelu wrote:
> > >
> > > > Hi,
> > > > About at the same IPC(semaphore,share memory,message) i make
references
> > > > too.Are more ways to made a inter process communication mechanism.
> > > > For example, in my applications i don't use semaphores and
messages.This
> > are
> > > > system functions available from the Kernel.
> > > > If you type "ipcs" you can see that mySQL don't use the system
functions
> > for
> > > > inter process communication.
> > > > My opinion is referencing at, if this CGI (about Shaun said early)
> > retrieve
> > > > huge data from RDBMS ,sure can create unbalanced processes.
> > > > For this reason i "said" it's "strongly recommended to setup MySQL
on
> > the
> > > > other host".
> > > > I have a bad experience with INFORMIX mounted on SCO running in a
dual
> > XEON
> > > > machine.And i think SCO it's more stable than Linux.
> > > > Of course who use mainframe don't must have any concern.
> > > >
> > > > Regards,
> > > > Gelu
> > > > _____________________________________________________
> > > > G.NET SOFTWARE COMPANY
> > > >
> > > > Permanent e-mail address : gg@stripped
> > > >                                           gelugogancea@stripped
> > > >
> > > > ----- Original Message -----
> > > > From: Svensson, B.A.T. (HKG) <B.A.T.Svensson@stripped>
> > > > To: Gelu <ggelu@stripped>
> > > > Cc: Jason Yates <jyates@stripped>;
> <mysql@stripped>
> > > > Sent: Wednesday, May 01, 2002 7:32 PM
> > > > Subject: RE: mysql speed concerns
> > > >
> > > >
> > > > > With IPC I mean Inter Process Communication - any process that
wants
> > > > > to talk with something else than with it self _needs_
> necessarily
IPC.
> > > > >
> > > > > Particularly an RDBMS needs IPC - otherwise it is useless.
> > > > >
> > > > > Anyhow. one way one might get faster performance could be to
> > > > > (if possible with the OS) move the swap space from disk into
> > > > > RAM instead. Or even better for speed and response time
> > > > > don't use swap at all! - but then you must know how much
> > > > > RAM memory your system will consume - otherwise you might
> > > > > get into big trouble. It is not impossible to estimate
> > > > > this, but tricky. One way could be to limit number of
> > > > > connections/jobs on the machine, and then assume the
> > > > > worst case scenario from this.
> > > > >
> > > > > > -----Original Message-----
> > > > > > From: Gelu
> > > > > > Sent: Wednesday, May 01, 2002 6:12 PM
> > > > > > To: Svensson, B.A.T. (HKG); Shaun Bramley
> > > > > > Cc: Jason Yates; mysql@stripped
> > > > > > Subject: Re: mysql speed concerns
> > > > > >
> > > > > >
> > > > > > Hi,
> > > > > > MySQL don't seems to use IPC.!?...Strange ..!?...And if
> Shaun
will
> > need
> > > > > > "real-time" response from RDBMS,in my opinion, is strongly
> > recomended to
> > > > set
> > > > > > up MySQL on the other host.
> > > > > > Regards,
> > > > > > Gelu
> > > > > >
> > > > > > _____________________________________________________
> > > > > > G.NET SOFTWARE COMPANY
> > > > > >
> > > > > > Permanent e-mail address : gg@stripped
> > > > > >                                          
> gelugogancea@stripped
> > > > > > ----- Original Message -----
> > > > > > From: Svensson, B.A.T. (HKG) <B.A.T.Svensson@stripped>
> > > > > > To: Shaun Bramley <s_bramley@stripped>
> > > > > > Cc: Jason Yates <jyates@stripped>;
<mysql@stripped>
> > > > > > Sent: Wednesday, May 01, 2002 6:20 PM
> > > > > > Subject: RE: mysql speed concerns
> > > > > >
> > > > > >
> > > > > > > An webserver uses mostly CPU resources while a DB uses
> I/O,
or?
> > > > > > >
> > > > > > > Separation between the webserver and the RDBMS also
> suggest
> > increased
> > > > > > > delays with IPC. I think one needs to find a balance
> somewhere
in
> > > > between,
> > > > > > > and I guess the only way to tell is to actually measure
> the
> > > > performance of
> > > > > > > the system  to see what's need to be adjusted in order
> to get
> > better
> > > > > > > performance.
> > > > > > >
> > > > > > > I would start out by collecting some stats from the
> RDBMS and
the
> > OS,
> > > > lets
> > > > > > > say during a weeks time or so,  and then just create
> dummy
> > instances
> > > > that
> > > > > > > perform the very same things and slowly increases the
> load in
> > steps in
> > > > > > > order to see what happens with the system.
> > > > > > >
> > > > > > > file://Anders
> > > > > > >
> > > > > > > > -----Original Message-----
> > > > > > > > From: Shaun Bramley
> [mailto:s_bramley@stripped]
> > > > > > > > Sent: Wednesday, May 01, 2002 5:13 PM
> > > > > > > > To: Jason Yates; mysql@stripped
> > > > > > > > Subject: Re: mysql speed concerns
> > > > > > > >
> > > > > > > >
> > > > > > > > Hello,
> > > > > > > >
> > > > > > > > The first thing that I would do would be to
> separate the DB
and
> > > > apache.
> > > > > > Set
> > > > > > > > the DB up so that it is on a box of it's own.  The
> system
cannot
> > > > serve
> > > > > > pages
> > > > > > > > if it is crunching through the db tables looking
> for
information
> > > > (and
> > > > > > vice
> > > > > > > > versa).
> > > > > > > >
> > > > > > > > Shaun
> > > > > > > >
> > > > >
> > > >
> > > >
> > >
> ---------------------------------------------------------------------
> > > > Before posting, please check:
> > > >    http://www.mysql.com/manual.php   (the manual)
> > > >    http://lists.mysql.com/           (the list archive)
> > > >
> > > > To request this thread, e-mail
> <mysql-thread107739@stripped>
> > > > To unsubscribe, e-mail
<mysql-unsubscribe-miranda=uranus.com@stripped
> > .com>
> > > > Trouble unsubscribing? Try:
http://lists.mysql.com/php/unsubscribe.php
> > > >
> > >
> > > --
> > >
> >
/*--------------------------------------------------------------------------
> > > * "It might look like I'm doing nothing, but at the cellular level
> > > * I'm really quite busy."
> > > *  ---anonymous
> > > *
> > > *
> > > * -----BEGIN GEEK CODE BLOCK-----
> > > * Version: 3.1
> > > * GCS/P/H/L/O d- s-:-- a26 C++(+++) UL+++$>++++ P++ L+++ !E W+++ N- o?
K-
> > > * w--(---) !O M+ !V PS+ PE- Y-- PGP- t+ 5-(--) X(-) R* tv b+ DI-- D--
G
> > e+>++
> > > * h+ r* y--
> > > * ------END GEEK CODE BLOCK------
> >
> --------------------------------------------------------------------------
> > */
> > >
> > >
> >
>

Thread
mysql speed concernsJason Yates1 May
  • Re: mysql speed concernsShaun Bramley1 May
    • RE: mysql speed concernsHKG)1 May
      • Re: mysql speed concernsGelu1 May
      • Re: mysql speed concernsGelu1 May
        • RE: mysql speed concernsHKG)1 May
          • Re: mysql speed concernsGelu1 May
            • Re: mysql speed concernsMichael Chang1 May
              • Re: mysql speed concernsGelu2 May
              • Re: mysql speed concernsGelu2 May
  • RE: mysql speed concernsJay Blanchard1 May
  • Re: mysql speed concernsJoseph Bueno1 May
  • Re: mysql speed concernsJeremy Zawodny2 May
Re: mysql speed concernsGelu2 May