System fail - help would be appreciated.

[BillCardiff]

Does the booting USB or the intended target USB save logs ?
After a power outage, the system is failing to reload. I am trying to figure out where to look to rebuild.
I created a new boot USB, and hunted down a new target USB. The system now reboots while creating the target when it's "loading the base OS" every time between 60 and 70 percent mark. Hoping I can use that to figure out what has failed in my system. Is it doing something specific at that point ? Maybe a RAM test, port check, hard drive check ? I know this isn't really helpful. Looking for any input or suggestions on what to check and where to look.

supermicro X8DTE
Xeon L5630 (pair)
96G RAM
12 hard drives across two flashed controllers

 

[Chris Moore]

This is one of the reasons why you should have a UPS, so there is no unplanned shutdown.
The boot media is likely borked. USB memory sticks are not reliable in this use and they fail often. You will probably need to reinstall FreeNAS and restore the system config database from the backup you made.

Sent from my SAMSUNG-SGH-I537 using Tapatalk

 

[BillCardiff]

This is one of the reasons why you should have a UPS, so there is no unplanned shutdown.
The boot media is likely borked. USB memory sticks are not reliable in this use and they fail often. You will probably need to reinstall FreeNAS and restore the system config database from the backup you made.

Sent from my SAMSUNG-SGH-I537 using Tapatalk

The UPS apparently has a dead battery - as of course it is plugged into a small UPS.
I am using found USB sticks - and name brand at that, but the system reboots itself as I said between 60 and 70 percent complete on the reloading.

 

[DrKK]

As Chris said, you need to REINSTALL FreeNAS on known-good USB sticks. The likely cause is the OS being bad. But depending on the cause of the power outage and the nature of the power outage, you could have other things wrong now. This will not harm your data, you'll simply have to import your pools.

Try that.

Unplanned power outages are bad for setups like yours.

 

[BillCardiff]

Wait. That is exactly what I am trying to do. HOWEVER - how do I know what a known good USB is ? I bought a Verbatim plus have a SanDisk and a Lexar - I am attempting to reinstall to these USB drives that were not in the system at the time of the power failure. I want to be clear that I have tried to install several times to different USB drives using other different USB drives as a source. I was hoping someone would reply with “check the logs at this location” as I can’t find any log information on the installing USB while the intended target USB is left in an unreadable state.

 

[DrKK]

"known-good" in this context means: either brand new, or surely known to be good. I think your new USB drives count as "known-good".

Your next step, I think, is to rule out one or both controllers being fried. Try booting the system with both controllers disconnected (yes, no drives), and see what happens. If you STILL can't boot, then you certainly have an OS install problem, or a more serious hardware problem.

 

[BillCardiff]

so an update.
I bought another set of Lexar 16G sticks, because that's my easy out. I pulled the pair of controller cards, and even pulled the hard drives slightly out of their cages so as to not draw power. Brand new USB pair. I got to 100 percent, and then the system power cycled itself.
I am thinking and now hoping that it's the power supply that has failed. I built this beast a couple or three years ago so there is no thoughts of a warranty of any sort. I am eyeing up a nice 750w with dual CPU leads, about $100 CAD. I can drive 45 minutes on the weekend or order one now and get it Monday, difference of about $1. Weekend it is. Here's to hoping it is just the power supply at this point.
The other thoughts for waiting for the weekend drive is that if it actually ISN'T the power supply and I need a whole new base, I can return the opened box for only a restocking fee. No shipping or fancy explanations required.

 

[Chris Moore]

I am thinking and now hoping that it's the power supply that has failed. I built this beast a couple or three years ago so there is no thoughts of a warranty of any sort.

Sounds reasonable. What kind of power supply do you have in there now?

 

[BillCardiff]

What kind of power supply do you have in there now?

650w corsair - semi modular. The planned 750 should give that little bit more head room.
My initial math said the ECC probably around 40W, 10 watts per hard drive x 12, 120W, each CPU 40 watts, 80W. I felt th 650 was within the math. 750 should give it a pile of room.

 

[BillCardiff]

And it's back... power supply. Replacement is an EVGA 750 BQ - has dual CPU support, 5 four pin peripheral, and a bunch of other connectors I don't care about. PIA rebuilding the case though, power supplies tend to weave themselves deep into the innards.