Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

I have a linux server whic is used as web/file/svn/sql server. It is accessed by 100 machines for storing their code files on this server(svn/code/mysql).

The OS is on a 80 GB HDD. The data is on another 500 GB HDD. Which is the best way to make a HA server?

1: Can I use rsync to mirror to a replicated server of the same config or should I add in another HDD and set up rsync to replicate the data. It's a basic server, so I dnot have hardware RAID 1(for mirroring).

2: Also if I set u rsync to run every 1 hour, will it increase server load to find the files which were changed in the past 1 hour and then copy to the backup server?

share|improve this question

3 Answers 3

I would stay away from rsync for this because, yes, it will increase the server load. Especially if you have lots of really big svn repos. Rsync has to traverse the entire file system and with the way svn is used for branches... Your talking LOTS of files. I would setup something like DRBD for replication to another node.

That being said, rsync may work for you and the load may be not have a significant noticeable impact on usability.

share|improve this answer

You have do differentiate what you want to achieve. You would use some replication technology to copy the data to another server to ensure that the data can still be accessed when the first server fails completely. You would however introduce RAID to protect against disk failure. You can create a software raid with mdadm without problems to increase availability, but you have to be aware that this does not protect you from server failure.

I'd suggest to do both: Configure software RAID1 and test the rsync approach on scheduled time intervals. There will be some performance impact (IO and CPU wise) but you can limit this by reading through the manpage and turning off features that might require excessive CPU like compression.

If you however want to have a "live" replication to a remote server have a look at drbd

share|improve this answer

You can use DRBD to replicate data to another machine. If you need just a backup, that's enough.

If you need a highly reliable service you can build a Linux HA cluster (heartbeat, Pacemaker, DRBD).

In both cases, I strongly recommend DRBD User's Gude.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.