I would like to be able to tell how much time my computers have spent running. I am not interested in how long they have run before needing to be rebooted—I tend to rotate from job to job, and they all spend more time shut down than running.

I realize that this raises the philosophical question, "What is a computer?". CPU, motherboard, HDD? It would be easy to treat a tower as a kind of "grandfather's axe—a new head, and three new handles, and as good as the day it was made". Looking online, it seemed that an easy measure would be the total number of hours the HDD had been running.

I also see that this information would be unlikely to be very useful as a predictor of how much longer a computer might last. . .

I am running Linux Mint Cinnamon 17.3, and looked at the Disk Usage Analyzer, which didn't seem to offer that information. Is there a command? And if so, would it give a global total, or only a total for the usage under the current distribution?

uptime is reset at boot so you need something like uptimed to track totals. This would only provide data since OS instalation. I have never seen any BIOS hour meter in a desktop PC which is why using HDD data is the next best approximation.

Hi,

I agree with michaelk that the use of 'hdparm' would be a good indicator for the system drive to show the hours for a System overall. That would be if that HDD was the original system drive.
Hope this helps.
Have fun & enjoy!

Some advanced bios makers might have logs for some clues. Some of the lights out boards might have logs. Unfortunately they can be cleared.

Someone said NASA bought 1000 of this special built computer for use on a mission. I asked the manager why would they buy 1000. He said that the computer maker tested them but NASA would put the most demanding tests. After many of them failed they'd select from the remaining ones assuming that they have a great chance of working. So, a computer that has been running longer may in fact be better.

Power_On_Hours parameter may be used as an indicator

But good luck restarting it if the power fails, at that point electrical and thermal shock will be your enemy

(Argh! Make that "scripting".)

If the system is brand new, you can keep track of the uptime and accumulate that with some scripting that runs at startup and shutdown. However, accounting for uptime that ends in a kernel panic or a power failure won't be as easy. You'd need to run a task that updates a system uptime accumulator every minute in, say, a cron job and use that to update an overall uptime accumulator stashed in a file somewhere.

Trivia: Many "large" computer systems used to have a actual "Hobbs meter" (you GA pilots will know what this is) will that recorded how long the system had been powered up and these would be rare on a system that wasn't installed in multiple 19" racks. I never asked the field service guys what they used this for. I assume it gave them an idea as to when certain preventive maintenance activities should be scheduled.

I have the impression that any utility I install will only record the hours following its installation?

As far as I know a correct statement and stated above without a hour meter as described above there is nothing that tracks real run time.

1 members found this post helpful.

Hi chaps,

Not quite true that there is nothing that tracks real run time.

I developed an hourmeter solution 23 years ago for exactly that purpose.

https://marcgloor.github.io/hourmeter.html

For instance, the current uptime and runtime of my cloud server:
I generally install this commmand "hm" right after setting-up either a workstation or a server.

Enjoy.

Cheers
Marc