HTPC update - separating backend from frontend

After using the Atom-based HTPC for a while, I noticed it start to struggle in high load situations - like playing back HD TV while recording multiple streams, commercial flagging, trying picture-in -picture, running disk checks. I decided to upgrade the box to something a bit more powerful, so there’s a bit more headroom.

So a while back I picked up an AsRock H77-M motherboard, with a cheap Pentium G2020 CPU. 4GB of RAM was sitting around idle after a previous desktop PC upgrade, so in that went. Two, 3GB drives were installed, still in a RAID-1 configuration. The OS was installed on an SSD that was left over. Graphics were handled by an Nvidia GT630 card, made by ASUS. I chose that one because of the passive cooler, and the fact that it is quite power efficient for what it does. A TP-Link wireless card was put in so it could see the network.

For the operating system, I went with Mythbuntu 12.04 again. Installation went pretty straightforward. No real problems, TV listing data was handled by the excellent shepherd - that just does its own thing, causing very little trouble.

Playback worked beautifully on this machine - it handled everything I could throw at it. However, my curiosity was getting the better of me, so I dug out the old Atom box and reinstalled mythbuntu, this time running just as a frontend. That made quite a bit of difference - without all the overhead of doing all the database tasks, flagging commercials and the like, it ran beautifully.

Since I had recently had some network cabling installed in the house, I tried running the new box as a backend only, up in the study, with just the Atom frontend in the lounge. Without any worries about wifi dropping out, the nice fast, stable wired network connection supplied data from the backend nicely.

It ran like that for a while, but then I started thinking that there are now three points of failure instead of one. Where previously there was just the Pentium backend/frontend to troubleshoot, there was now also the frontend box, plus the router that could cause problems. Eventually I put the frontend box back away.

I’m not sure what made me do it, but soon after getting a new TV (LCD! HDMI inputs! No more fooling around with xorg.conf and custom modelines and VGA to Component transcoders!) I wanted to clear out a bit of clutter from the lounge room. Moving the big box meant I could get rid of a whole heap of things from the lounge - the UPS, the transcoder (not needed any more), the aerial connection with splitter and the external TV tuners. All of the bulky component cables are no longer needed, either, replaced by a single HDMI cable.

So up to the study the box went, along with all the tuners and UPS, but they have been tucked away behind the desk, out of sight. The lounge room now just has a power cable to the frontend box, a network cable, an optical cable to the receiver, and the HDMI cable to the TV.

For now, I’m just rebooting the router every few weeks, or whenever there is some sort of issue like PCs not being able to access the internet. A lot of data is now going across it with TV viewing.

The backend box was able to have the graphics card removed from it, as it is now running headless. I only connect to the machine via ssh. That has saved a bit of power consumption. Another change made to the backend box has been to add a second SSD, to run the OS off a RAID-1 array. This way, a drive failure will not stop the machine. It will run in a degraded state until a new one can be installed.

I did the same with the frontend box - a couple of old laptop drives were pressed into service for that. Both are pretty elderly, but since they are in RAID-1, unless they both die at the same time, it should be safe. And really, it’s only a frontend, it wouldn’t be too much of a hassle to rebuild it from scratch if needed. I’m considering replacing those drives with a couple of USB sticks, to further reduce power consumption.

For the backend box, disk usage is sitting at around 65%. I’ve been copying some of my DVD library to the server, just to avoid dealing with physical discs. I don’t actually have a DVD drive in the frontend, anyway. If space gets tight, I might need to expand storage. I’m toying with getting a third 3TB drive, and converting the system to either RAID-5, or going a bit radical and putting ZFS on Linux back on it. My only concern with that is that it can have issues with fragmentation, and my PC doesn’t use ECC RAM. A lot of memory is also used for caching purposes with ZFS.

One thing I am wondering about with my current XFS file system is to do with fragmentation. I currently have a cron script to defragment it nearly every morning. Not sure what the deal is with RAID-5, and how it deals with fragmentation. Might just have to try it and see, when I get around to it. Might just go with ext4 for the file system. Outright performance isn’t so much of an issue for this machine - video recording and playback isn’t pushing it too much.

Adding a larger drive to a RAID array - ZFS

If you have waded through my previous post about adding a larger drive to an lvm volume running on an mdadm RAID array, you'll have seen it's a bit of a fiddly, complex process. I have been fiddling around a bit with ZFS on Linux on my old desktop machine (Core 2 Quad Q9300, 8GB RAM). I had a couple of old 1TB drives, and added the 2TB drive that I removed from the home theatre PC, mentioned in the previous post. Actually, I had put the 3TB drive in this old box initially, intending it to replace the HTPC. But it's a bit power hungry, using around 90W at idle, and I am having some niggling playback issues with it. But that's another story. So the 3TB WD Red drive gets replaced with the old Samsung 2TB.

With the three drives, I tried creating a ZFS raidz array, basically equivalent to RAID 5. It runs pretty well. Swapping the drives over (from 3TB to 2TB) gave me an opportunity to do the ZFS version of a drive swap.

It was a bit of an anticlimax, really. There was one command, the syntax from the man page is below:

zpool replace [-f] pool device [new_device]

In my case it was something like

sudo zpool replace tank /dev/disk/by-id/scsi-SATA...(the drive removed) /dev/disk/by-id/scsi-SATA....(the 2TB drive)

That started off the resilvering process (resynchronising the data across the drives). If I then were to replace each other drive, one at a time, with the above process, it would resize the array to match the new drive sizes.

The beauty of ZFS is it simplifies things so much. It is the RAID management and file system all in one.

I would have liked to run it in the HTPC, but ZFS is quite resource-hungry, in terms of processing power and memory. Since my HTPC just has a little Atom 330 chip and is maxed out with 4GB of memory (of which only 3GB is visible), I would be staying with mdadm and lvm. There is virtually no noticeable performance penalty with it. Oh well, going through all those steps is not exactly a frequent task.