My trusty home server decided to finally give itself up after losing it’s operating system drive, the root mount about three month ago. Yes. My server was running off RAM for the past three months where I was really really busy with my school work – my Honour Year Project – to be exact. As such, I have lost my last blog post on the labeling of ZFS volumes in FreeNAS. But I’m very grateful, it managed to stay alive until yesterday, one day after I gave my last undergraduate academic presentation for my Honour Year Project.
This time round, I’ve decided to run this wordpress app on my edge hypervisor box with a guest OS of FreeBSD. Setting up a FreeBSD server takes time but it’s quite a rewarding experience, well, this might be due to seeing all the codes compile for the necessary services.
The Hard Drive that died was a 2004 Seagate 80GB SATA-I hard drive that was moved across multiple computers. I’ve already knew about the loss of the mount about 3 months ago, but Linux being Linux, it’s quite awesome where everything is still kept in RAM and able to run my wordpress and torrent box like nothing had went wrong. But it seems like time is up. I’ve received the death alert by my remote monitoring system yesterday afternoon.
The death was timely too, I have a upgrade plan to move my storage server to a complete FreeNAS system, using ZFS and hard disks in mirror mode with ZIL and L2ARC with SLC and MLC SSDs. This change in storage system will give me some storage performance boost that is capable of providing iSCSI for my hypervisors, and CIFS/NFS for my storage client machines over my home network.
Just yesterday, I was caught in a very interesting discussion of scalable storage/database reliability between two experienced techies of SOC. They were discussing about the move of storage towards object storage system such as Ceph, and leveraging on various techniques like load-balancing and distributed-hash-tables (DHT) to really grow and scale in an incredible way. A way that never crossed my mind for my Honour Year Project. Maybe after graduation and moving around my computing resources, I could experiment on that technique too.
In the meantime, I still have to tie some loose ends of my HYP and then my last two papers of my Undergraduate life. Yes. I’m finally graduating. 🙂