On Fri, 19 Jun 2020 at 10:43, Justin Moore <justin.nonwork(a)gmail.com&gt; wrote: Start keeping a log. For years I was going in to work 7 days a week to babysit an unreliable system used to capture "irreplacable" data. From the logs I determined that the system never failed on Sunday, so I went from a 7-day work week to 6 days. "last" should show if you had an orderly shutdown or a crash. There is a good reason that linux distros include a memory test -- use it. Sometimes reseating memory cards is enough to get past a test failure. After memory common old PC random crashes are due to bad cables, connectors, or power supplies. Particularly if you live in a hot, humid, area (or take your PC to sea) you may find green corrosion on some connections. You can buy "contact enhancer" that is really helpful for seagoing or ocean adjacent gear. At my work we generally had a stack of old PC's "all made out of ticky tacky. And they all look just the same…", so I could swap cables and power supplies to keep a few running. Before retiring I managed a 7x24 industrial PC used to control a satellite dish. Twice the system started having random crashes. Inspection of the system board revealed bulging capacitors. <https://en.wikipedia.org/wiki/Capacitor_plague> I bought a new MB and replaced the bad capacitors on the original. Then the new MB failed the same way, so when I retired it was running on the repaired board. That MB supported a range of CPU clock speeds. Normally I run 7x24 systems at the lowest clock speed for longevity, but we were using time consuming software decryption so the system used the fastest available CPU. -- George N. White III