I moved my Owncloud instance over to a Raspberry PI 2 a month or so ago. A couple weeks ago we had an extended power outage and it appears it went down hard and corrupted the disk. I fsck'ed it and it started going again but went down a couple days later and the drive was corrupt so I fsck'ed it again and it ran for a couple more days then went down again. I figured the original power outage messed it up so I reformatted and started clean. I ran slackpkg-update and rebooted afterward and it didn't come back up and again I had to fsck the drive.

I'm wondering if I have a failing microsd? What brand do you run? How can I test it? Do you leave your / (root) on the microsd or do you run it on a usb flash drive? What are the best practices?

I had a similar situation and the SD card failed, went completely DOA. Root cause was an under-powered power supply.

Chuck your SD card and get a good one. I've been up now for 40 days.

I use a 8GB SanDisk UHS-I

Hmm, I did swap the PS for a small powered USB hub so I could power both Pi's off of one plug. I suppose high utilization could be part of the problem, loading the ps down.

RasPi's can be power sensitive and not all power supplies are created equal. I've had all sorts of problems with "generic" micro-usb power supplies that only push out 500ma, when I changed to a 1amp I had no further problems.

This is supposed to be 1A per port but is a generic no name. I can install clean then do the slackpkg ugrade-all and reboot and it will die. I bet its corrupting the drive as its writing the installs. I'm working around that by resyncing my source so everything is up to date during the initial setup then there will be less to upgrade.
Will try a different ps tonight.

I gave up on the generics. You cannot trust their labels. They target phone charging.

I finally went with a 2.2A switcher from MCM. Guaranteed to work all needs up through the Pi 2 model B. Now if something fails, it's not due to dodgy power supplies.

I need something with multiple ports, I'm running my PI's inside a small network cabinet.

I need something with multiple ports, I'm running my PI's inside a small network cabinet.

Are you all using microsd for root or are you moving it to a usb flash drive and just booting off the microsd?

Depends on what you want to do. I just have root, var, usr, home etc on the SD and application specific data filesystems (i.e. apache docroot, motion directories) on separate flash thumb drives plugged into a powered hub.

well crap, now I'm fighting multiple issues. arm-current is borked, it gives a bunch of ldconfig errors and doesn't find eth0

The very best thing you can do for any rpi/rpi2 is get / and swap off that damn sdcard and onto a proper hard drive. Mine run rock solid after that.

Also, after any write to /boot, the only thing left on the sdcard, sync and unmount it before shutdown/reboot. Not convinced that the card itself gets long enough to "settle" before the system reboots itself.

Edit...
Also (pt2) all the slackware installer images start at sector 32, As I understand it, this puts the MBR far too close to the FAT of mmcbmk0p1 meaning that any write to the FAT is potentially within the same stripe block. Note that all sdcards now have their partitions starting at 8192.

If you mount the .img file (losetup and offset) you can copy the contents of the image to your sdcard with P1 starting at 8192, this also seems to help.

By proper hdd do you mean a spinning disk or is a USB flash sufficient. I don't have any hdd's running now, my data is on a 128G USB flash.

If you have more than 1 warning from ldconfig, it means you did not update your aaa_elflibs package. If you have one warning then it's expected since one package made it into that batch prematurely.
See http://www.linuxquestions.org/questi...ts-4175541253/
There's a new set of updates ready to go out when x86 goes. You should update to the new aaa_elflibs package in that update as well.

Can you expand on what you mean by "find eth0". If you mean that there is no eth0 interface present using then that normally means that if the network driver is compiled a a module then the Kernel module is missing, or udev failed to load it for some reason. Check the log messages and dmesg. There are no such issues on all of the four varieties of build machines that built -current, but I don't know a lot about the customisations for the Raspberry Pi stuff.

If config didn't list eth0. I suppose it was just bad timing, I had resynced my arm-current then did a clean install and saw all the ldconfig messages and no eth0 so I assumed related but maybe not. I updated Exaga's driver/firmware packages and its working now. I think the logs have wrapped already to go back and see any details.

ok, def seems related to the microsd
I've been running 10 days now since moving / from microsd to another usb flash drive, so far so good. Did she slackpkg update a coupld days ago with maybe a dozen new packages. No crash yet.

I have not imported my calendar back into the database or swapped the power supply, I'm trying to test one change at a time.