On 30/08/2015 17:24, Daniel Frey wrote: > On 08/30/2015 06:24 AM, Michel Catudal wrote: >> >> As for shutdowns there are several arguments for and against. What often >> kills electronic is the shock between hot and cold so there is an >> argument about keeping the system on. >> Whether it is always safe to keep the computer on all the time remains >> to be proven. > > Recently I've had to help someone migrate off of a failed computer. This > computer was old (I had to find an IDE adapter to recover some files) > from late 90s/early 00s. > > Some time ago I told him to have it running all the time, mostly because > of age. So he kept it running nonstop and literally a week or two ago > shut it down as he was getting new flooring installed. He called me > after hooking it back up again as it wouldn't start. I went over to > check and the motherboard finally failed. He hadn't powered it off in > 4-5 years. > > For myself I use a smart power bar and suspend my PC when not in use. > This caused me all sorts of grief with systemd hanging on shutdown after > a suspend, ultimately causing my RAID array to be rebuilt on every > reboot/shutdown and so I've finally abandoned it and am running openrc > again. > > The only thing about using suspend is that if the PC is in a sleep state > it won't wake up and shut down when the power goes out. This just > happened to me yesterday (big wind storm here.)
One of the reasons sysadmins have old servers out there that still have huge uptimes, is that we dare not switch them off. We don't know if the drives will spin up again from cold! Technically, we should do a power down test every 6 months or so, but that turns out not to be a yes/no test in real life; it's a yes/destroy test and no-one wants to make a decision either way. So we all sit in limbo and wait for some exterior event to decide for us (like black-outs) Sad, init? -- Alan McKinnon alan.mckin...@gmail.com
