New Year, New Tape!

Well, I finally got it, I flipped enough switches, and "plugged" and "unplugged" enough connections to get my NAS server to see the tape library! Specifically I think I kept stumbling on the fact that I wasn't trying to emulate a device, just trying to connect to one. This means that I don't need to load the target module(qla2x00tgt) for the fiber card.

Loading the target module blocked me in a couple of ways. First loading the module switched the fiber card from "initiator" to "target" mode, which cause the device to look for incoming connections instead of reaching out for other targets. Second it gave the impression that things should just "work" once you have them setup. This impression was re-doubled each time I tried to google what was wrong, as the blogs, articles, and forum posts all were trying to do the same thing. When things didn't just "work" there was no guidance on how to troubleshoot, just instructions to recheck all the boxes.

This lead to a 3-ish months* worth of failed attempts, frustration, stepping away, repeat. Finally after much thought I realised that I might be shooting myself in the foot, and started trying without adding the target module. Once I did that, everything fell into place(kinda). The device started showing in the /dev/ folder, I could run the commands in the original post and have them return data.

Once I got it to return data, I started attempting to write to the tapes. This has gone ...."ok" is probably the word I'm looking for. I'm running into issues due to the size of data that I'm trying to backup, along with the age of the tools that are written for this type of stuff. This may sound like a cop-out, but both the command tar and the gzip command are written in a single threaded way. When you're dealing with ~8Tb of data, running it through either of those commands, means that you have DAYS of waiting for the commands to complete.

I've tried a couple of time to write to the tapes, and had 1 successful write that took 3 tapes. Though when I tried to read those tape back, it failed to read back, but I only gave it 1 shot. I've now moved on to just backing up the file into a tar.gz file, with the intent of testing the drive, and validating I can read/write correctly to the tapes. 

Once I confirm that I can write/read to the tapes the way I expect them too, I will move onto dealing with the size of the dataset. I already have plans to try pigz for multi-threaded compression, but I don't yet have an answer to tar just being slow in general. I also don't have an answer to "how does this go across tapes?" since tar can't do compression and multi-tapes at the same time. 

One step at a time though, I'll worry about the size of the data set once I get a good feel for how to best use the tapes, or at least how to best use the tapes for what I want out of them. In the mean time, its a waiting game. Waiting to make a backup, waiting to write the backup, waiting to read the backup, and finally I will wait to test the backup and make sure it came out ok.

* you might wonder why it took 3 month, I will gently refer you to the posting date of this article, and look back at your last 3 months(if your in america) and ask when you had time to do debugging. If your not in america, the word I have for you is "Holidays".