We all get reminded about backups regularly, and many seem to forget until the problem occurs so I thought it was a timely reminder to do backups, and not just one, many where possible.
I had my primary hard drive fail on Friday evening. It’s a solid state and like anything, they aren’t immune to failing. It’s a 2015 model so it would be going on about 3.5 years old, just outside of warranty. I’m not going to question whether it should be covered, 3 years warranty is about the usual you’d expect these days so having a hard drive fail on 3.5 years isn’t something I can blame on the manufacturer. Sometimes these things last 10 years, sometimes not.
My backup strategy
Fortunately, I’ve always taken backups seriously. I backup to multiple locations, namely:
- Time machine backup for incremental changes – This backs up my hard drive on an ongoing basis ensuring that if my harddrive fails at any point, I’m unlikely to lose more than an hour or two.
- Online backup for offsite backup – This ensures that if my house burns down with both my laptop and hard drive, I still have a backup.
- Drive clone – This ensures that if I have urgent work to conduct, I can boot off the clone, and start working within a matter of minutes by accessing the time machine to get anything I need.
Three tiers of backup may seem excessive but I would recommend at least one and two because this covers you for hard drive failure and for acts of nature that might destroy your house.
How do you know when your drive is going to fail?
In some cases you don’t. You start your machine and nothing happens.
In my cases, there were some warning signs. I had some suspicions things were going pear-shaped but I knew I had the backups. For me, the warning signs were:
- Machine slowing down – OSX is normally a good performer but I could see performance degrading more than usual. Performance may degrade during heavy usage but when it degrades when you have CPU and RAM to spare (I use iStat menu), you start to get concerned.
- Kernel panics – This was the main warning sign for me. Once in a while you might get a kernel panic on OSX, once a year, once a month would be really concerning and only likely if you’re using betas but generally Mac is pretty reliable. Kernel panics for those who don’t know, are the OSX equivalent of the blue screen of death on Windows. It requires the system to become unstable which in Unix land is unusual, and more often than not, it’s the sign of a hardware problem. When OSX first starting using dual CPU’s, these were a little more common as the OS battled to switch GPU’s but in the last 5 years, it doesn’t occur often. When you get 3 in one week, you start getting really worried that something is going wrong.
- Corrupt application or files – applications don’t get corrupt often, mailboxes don’t get corrupt often, files don’t get corrupt often, and definitely not in close proximity from a timing perspective. I had both occur within a week. When they both happen, you’ve got reason to get worried something is going seriously wrong.
- Can’t reformat drive – In an attempt to check whether there was a problem with the OS, I tried reinstalling. It failed every time I tried. Warning sign number 4
In combination, these gave me the clear indication that I was going to be replacing my drive. Knowing I had backups mean I could continue as needed, waiting for the failure to occur and time machine meant that even file corruptions could be rectified if I restored my landscape and discovered a bad file.
If you haven’t worked it out already, and you don’t have a backup, get one. You can buy a hard drive for under $100 which includes backup software. If you’re on Mac, OSX has backup software built-in that will allow you to back up the machine easily so there is no excuse not to back up.
Ultimately, when it comes to backups, it’s not a case of if, but when. I’ve had two failures in 10 years, if you haven’t had one, it’s only a matter of time.