Client is running a business portal that have approx. 300-500 MB data uploads per day. There was NFS Storage server which had 500 GB drive which was running out of space. This Server was also running lighttpd web server to provide access for users to view these files.
No Failover. (Except offline backup)
When ever server gets done, Users are unable to upload data.
Users are unable to access their data because lighttpd was also running on same server.
This client was looking for a cheap solution which can work for next 10 years for them.
Solution we proposed and implemented :
As always we go for Open source solution.
Two dedicated servers
600 GB SAS 15K RPM X 6 on RAID 5 on each server.
Implemented DRBD on two nodes for data replication on over network.
heartbeat for high availability
lighttpd, nfs was also installed on both servers
Virtual IPs was assigned on LAN and WAN interfaces for smooth fail over.
We chosed two servers because of fail-over
600 GB x 5 on raid 5 means we get 3 TB storage which is enough for next 15 years.
lighttpd was installed on both servers to serve static requests.
NFS was installed to export share to apache cluster.
Virtual IPs is used by heartbeat service for failover.
NFS – is a Network File Sharing service.
Lighttpd – is a http web server.
DRBD – is a distributed storage system on linux.
In next phase, i will write a brief HowTo for this whole implementation.
If you know the better implementation, please do share with us. We are keen to know that in comments or suggest us at tips-@-hackersgarage-com
Stay tune for more updates!