any reasons that filesystems become corrupt?

 

this could have been a much bigger mess if it was my media or home photo drives

Unclean shutdowns due to power cuts or system crashes are usually the number one reason.

That's a likely cause then. As stated when embryo was stalling the server wouldn't actually reboot.

 

not even talking safe shutdown (which I wasn't able to find up to date doco on)

 

im talking it accepting a shutdown, power off, reboot cmd but never actually rebooting

 

lsof commands to find stuck open files etc would just never return anything to even diagnose why.

any suggestions or calculators you know of in terms of a UPS ?

 

currently running 7 drives, and would want about 15-20 up time then shudown cleanly

3 hours ago, kingy444 said:

currently running 7 drives, and would want about 15-20 up time then shudown cleanly

Very large... 1500VA probably.

You would want a UPS with a total possible runtime of close to an hour to comfortably run for 20 minutes without draining the batteries to the point of stressing them. You really don't want to run below 50% of total battery capacity, it shortens the lifetime of the batteries badly.

What sort of runtime would you expect from a 1200? Need to find that happy medium between killing the ups and up time in an outage.

if I have to cut it short to save some $$$ then I will.

 

i was currently looking at the APC BR1200GI ?

6 hours ago, kingy444 said:

Need to find that happy medium between killing the ups and up time in an outage.

I know it sounds nice to be able to say my server stayed up for X minutes during an outage, but, honestly, what happens during an outage that using your server is so important? I know personally when the power goes out for more than a minute or two, it's going to be down for probably hours, so I spend the server uptime running around shutting various clients down and making sure things get saved. Maximum I've really NEEDED the server to stay running would probably be 5 minutes or so. Most stuff is set to automatically start the shutdown process after 3 minutes without power. If you do regularly run the USP batteries completely flat, you will need to replace them more often, but that would be cheaper than a huge UPS.

 

If you honestly have a real need for extended device uptime in an outage on a limited budget, I suggest rolling your own UPS with a decent quality inverter/charger linked with 12V wet deep cycle batteries. You can get many hours of runtime, and if you need more time, a pair of jumper cables to a running car every once in a while will give you power until you run out of fuel.

Agree with jonathan, I have my UPS set to shutdown after 5 minutes, here if the power is out for that long it will usually be out for much longer.

 

As to give an idea on runtime, I have two APC BR900-Gi, these are the currently reported runtimes (which are usually a little optimistic)

 

1 - Load 200W - runtime 28 minutes

2 - Load 80W - runtime 54 minutes

Ok, so it sounds like the 1200 would be enough, maybe even the 900 might be ok.

 

how do yo calculuate the Watt usage of your server?

 

from what your saying 5 is really going to be the   Longest I need to keep it up. And then start shutdown tasks.

7 hours ago, kingy444 said:

how do yo calculuate the Watt usage of your server?

 

Get a Kill-a-watt or similar.

so as suggested, just impletementd a cache pool as explained here http://lime-technology.com/wiki/index.php/UnRAID_Manual_6#Creating_a_Cache_Pool

 

I just want to confirm process i undertook and ensure that the pool is valid.

 

So i precleared one 2TB WD Black, and that went fine. I assigned as cache drive one and formatted to btrfs. I went to preclear 2TB WD number 2, and appears to be DOA.

So i put my old 1TB WD back in there, and added it to the pool. This drive was previously formatted at reinfrs but i wasnt prompted to format the drive. I assume UNRAID did the formatting for me ?

 

Yes, any device added to a pool will be btrfs formatted by unRAID.