I bought an used Macbook Pro Mid 2012 and it has performance issues. Basically it hangs from time to time even when its not under heavy load. I started to inspect the possible causes and I found out the SSD unsafe shutdown count is 18 183 what is a really huge number, my co-workers laptop has like 8-20.

enter image description here

The idea I came up with is that maybe my ssd lost power randomly, what the system can handle (with simply waiting until the ssd will be available again) but the system will hang until it the ssd comes back.

Could this be the cause of the performance regression?

  • Doesn't the tool say that 18,183 is 96% good? Let's assume you are about to drop to 95% - then this tool is rating that drive at 18,184 / 0.05 = 363,680 unsafe shutdowns? I would look at the system logs surrounding a pause/freeze - if in fact the IO stopped, there would be enormous errors and timeouts logged. You could also keep a log of the pauses and see if they correlate with that number increasing. My intuition tells me it's not related, but I've also not used the tool you mention. – bmike Feb 19 '16 at 14:19
  • The percentage is an estimation how many times the event can happen before the ssd die. So it means this event can happen roughly ~ 345477 times before the ssd die for sure. What is the easiest way to get the IO logs from a Mac? Where is it placed? – NoNameProvided Feb 19 '16 at 14:38
  • The console.app has those details. I've always taken those "lifetime estimates" as just that - back of the envelope guesses as opposed to deterministic facts. It's good you're watching, but I doubt if you had 10 drives each would stop working between 345460 and 345480 - basically this test is very precise, but not very accurate. – bmike Feb 19 '16 at 14:43

You must log in to answer this question.

Browse other questions tagged .