Re: Critical mysql bug = remotely changing from PURE64 to 32-bit debian.. possible?

2005-04-11 Thread Pete Harlan
On Sat, Apr 09, 2005 at 07:29:53AM -0700, Jeremy Gray wrote:
 
 Is it possible to go from a Sarge Pure64 installation to a 32-bit debian 
 installation remotely?
 
 (I have KVM over IP access if that helps in the process any.. but no 
 CD-ROM on that server)
 
 If its possible could someone please tell me the steps to do it or point 
 me to a url that covers it?
 
 Thanks!
 
 See below for why I think I need to switch to 32 bit

Without knowing exactly what's causing his problems, I wouldn't worry
about them too much unless you see them on your own setup.  There are
surely plenty of cases where people have run MySQL successfully in
64-bit mode, on similar hardware and software configuration as his,
without problems.

--Pete


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Critical mysql bug = remotely changing from PURE64 to 32-bit debian.. possible?

2005-04-09 Thread Jeremy Gray
Is it possible to go from a Sarge Pure64 installation to a 32-bit debian 
installation remotely?

(I have KVM over IP access if that helps in the process any.. but no 
CD-ROM on that server)

If its possible could someone please tell me the steps to do it or point 
me to a url that covers it?

Thanks!
See below for why I think I need to switch to 32 bit
-
The reason I want to drop down to 32bit is because of the following 
critical MySQL bug/problem referenced below.
This server is going to be a mission critical production mysql database, 
so we don't want to take any chances.

@ http://bugs.mysql.com/bug.php?id=8555
See post...

*[11 Mar 11:19pm] Michel Buijsman*
We're now running in 32 bit mode on the amd64, and that seems to work fine. One
of the servers is now running as a replication slave with live data and it's
actually replicating which is more than I got when running a 64 bit OS.
I've done some head-to-head speed tests (2 identical servers, one with a 64 bit
install and one with a regular 32 bit debian, both running mysql 4.1.10 with
identical configs) and the 32 bit version is about 10% slower.

See also: http://bugs.mysql.com/bug.php?id=3483

--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]


Re: Critical mysql bug = remotely changing from PURE64 to 32-bit debian.. possible?

2005-04-09 Thread Paul Brook
On Saturday 09 April 2005 15:29, Jeremy Gray wrote:
 Is it possible to go from a Sarge Pure64 installation to a 32-bit debian
 installation remotely?

 (I have KVM over IP access if that helps in the process any.. but no
 CD-ROM on that server)

 If its possible could someone please tell me the steps to do it or point
 me to a url that covers it?

Use [c]debootstrap

http://www.debian.org/releases/stable/i386/ch-preparing.en.html#s-linux-upgrade

Paul


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Re: Critical mysql bug = remotely changing from PURE64 to 32-bit debian.. possible?

2005-04-09 Thread Frederik Schueler
Hello,

is this bug reproducible only if the highly experimental gcc-3.4/gcc-4.0
branch is used, as stated in the quoted mysql-bts bugs, or does this 
happen using a standard debian-pure64 (gcc-3.3 based) installation too?

A quote from mysql bug #3483:
 I've tried 4.0.23, 4.1.9, 4.1.10 on two dual opterons with 8 gig ram
 running Debian 3.1 pure64 gcc3.4. Tried kernels 2.6.11-rc1-mm1 and 
 2.6.11-rc2.

the gcc-3.4 (now gcc-4.0) branch is based on an unreleased, experimental
compiler version (http://gcc.gnu.org/gcc-4.0/). The code is likely to 
break and _not intended for production_. 

Again: if you want to run a production system with 64bit userland, pick
debian-pure64, and not the experimental gcc-3.4/gcc-4.0 based branch.


Kind regards
Frederik Schueler

-- 
ENOSIG


signature.asc
Description: Digital signature


Re: Critical mysql bug = remotely changing from PURE64 to 32-bit debian.. possible?

2005-04-09 Thread Jeremy Gray
Frederik Schueler wrote:
Hello,
is this bug reproducible only if the highly experimental gcc-3.4/gcc-4.0
branch is used, as stated in the quoted mysql-bts bugs, or does this 
happen using a standard debian-pure64 (gcc-3.3 based) installation too?

A quote from mysql bug #3483:
 

I've tried 4.0.23, 4.1.9, 4.1.10 on two dual opterons with 8 gig ram
running Debian 3.1 pure64 gcc3.4. Tried kernels 2.6.11-rc1-mm1 and 
2.6.11-rc2.
   

the gcc-3.4 (now gcc-4.0) branch is based on an unreleased, experimental
compiler version (http://gcc.gnu.org/gcc-4.0/). The code is likely to 
break and _not intended for production_. 

Again: if you want to run a production system with 64bit userland, pick
debian-pure64, and not the experimental gcc-3.4/gcc-4.0 based branch.
Kind regards
Frederik Schueler
 

I haven't tried running it yet myself, but I am just on the 
Debian-Pure64 branch.

I did see the notices that gcc-3.4 is experimental.
I first came across the notice of this bug at Joe Stumps page:
http://www.joestump.net/193251484/running-apache-mysql-php-on-debian-amd64.html
Not sure what version he was running, but he put a comment in on the bug 
too.

So it might be fine in Pure64 mode then?
Thanks


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]