LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

RE: Network Block Device w/ RAID for Content Replication (Was: Re: load

To: "lvs-users@xxxxxxxxxxxxxxxxxxxxxx" <lvs-users@xxxxxxxxxxxxxxxxxxxxxx>
Subject: RE: Network Block Device w/ RAID for Content Replication (Was: Re: load balance web site)
From: Nicolas Huillard <nhuillard@xxxxxx>
Date: Sat, 3 Jun 2000 22:17:13 +0200
Your setup seems very interesting.
Just one question below...

-----Message d'origine-----
De:     Atif Ghaffar [SMTP:atif@xxxxxxxxxxxx]
Date:   samedi 3 juin 2000 15:23
À:      Andi Hechtbauer
Cc:     tc lewis; Joseph Mack; pu@xxxxxxxxxx; lvs-users@xxxxxxxxxxxxxxxxxxxxxx
Objet:  Re: Network Block Device w/ RAID for Content Replication (Was: Re: load 
balance web site)

... 
There are 6 frontend servers that are running apache, php3 modperl etc.
each with 2 9 gb disk.
the first 9gb disk is for system (/boot, / etc)
the second 9gb disk is /www this disk is running software raid 1
with the nbd that is mounted from the main file servers.

[Nicolas Huillard]  I don't understand how the frontend server's second disk 
are managed :
* do they see a RAID 1 array (can you remind me what is RAID 1 ?) with a local 
9GB SCSI disk and a distant nbd disk ? In this case, the back-end servers will 
have to manage 6 disk for the 6 front-ends ?
* or is it the back-ends that have a RAID array, consisting on a local disk and 
6 distant nbd disks (one for each front-end), in order for them to see the 
content, read-only, on their local disk ?

Last question : how do you manage MySQL fail-over between the two back-ends ?

TIA,

Nicolas Huillard


<Prev in Thread] Current Thread [Next in Thread>