A new data architecture on NAS DELL MD3000i to accomplish replicating data on virtual machines

abril 5, 2011


No matter if you use virtualization or not your goal must be to maintain data as secure and stable as posible. To accomplish that you must use replicated data on RAID disks and choice between RAID 5 or RAID 6 levels. Now is no time to detail issues about them but both have a significant caracteristic: you loose all your data if 2 disk or more crash inside the group of disks. At our company we use a NAS Dell MD3000i with 15 SAS disks that serves iSCSI targets to 3/4 servers running VMWare ESXi 4. At first configuration, we left only one disk as Hot Spare and create 3/4 groups of disks containing RAID 5/6 data. This was done very fast because we had no time to think about a special configuration.

In first graphic you can see that VM’s where contained inside RAID 5/6 group of disks and our point of failure was expecting to loose only on disk in every RAID because another disk failure would drive us to a big headache, loosing all our machines contained in. Other issue was we highly depended on Dell Support to replace disks because we only have one hot spare.

We decided to change all. We backed up all critical data. Then move all VM’s to a iSCSI connected and VMWare certified Iomega Storcenter ix4-200d and let play users with another iSCSI NAS. Our first operation was to delete all virtual disks and leave 5 disks as hot spare. Then we divided the rest of 10 disks into 2 groups of RAID6, each one of 5 disks. We created OVF exported VM’s and then we put then inside the first group of disks, connecting that group to all of out ESXi servers. We created new VM’s importing that OVF backup using 30min for each VM. Finally for each VM with critical data we create one virtual disk on group #1 and other on group #2. Once the VM was up and running, using Microsoft iSCSI Initiator 2.08, we created RAID1 software disks using pair of disks, one from group#1 and other from group#2. At that point we increased 100% of data safe because now we needed 4 disks failures to loose all our data, 2 disks on RAID6 at group#1 and other 2 disks on RAID6 at group#2.

In second graphic you can see every VM is connected to iSCSI MD3000i data using two channels, one for the first disk and other for the second disk at RAID1 software.

Anuncios

Responder

Por favor, inicia sesión con uno de estos métodos para publicar tu comentario:

Logo de WordPress.com

Estás comentando usando tu cuenta de WordPress.com. Cerrar sesión / Cambiar )

Imagen de Twitter

Estás comentando usando tu cuenta de Twitter. Cerrar sesión / Cambiar )

Foto de Facebook

Estás comentando usando tu cuenta de Facebook. Cerrar sesión / Cambiar )

Google+ photo

Estás comentando usando tu cuenta de Google+. Cerrar sesión / Cambiar )

Conectando a %s

A %d blogueros les gusta esto: