system freeze[solved]

[maverick0815]

Recently my primary harddisk died, so I replaced it with a brand new drive. So currently I have two 2TB drives installed. The new drive holds a swap partition, a system partition and a data partition.
The second drive is only for data.
I did a fresh install last weekend and so far aside from the mythtv-dying problem everything seemed fine.
However I'm now experiencing system-freezes at least once a day. I checked around the logs and find things like this:

Jul  5 17:02:45 dcerouter kernel: [36587.564908] mdadm: sending ioctl 1261 to a partition!
Jul  5 17:14:55 dcerouter kernel: [37317.631345] ata5: hard resetting link
Jul  5 17:14:55 dcerouter kernel: [37317.631347] ata5: nv: skipping hardreset on occupied port
Jul  5 17:14:56 dcerouter kernel: [37318.100030] ata5: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
Jul  5 17:14:56 dcerouter kernel: [37318.132515] ata5.00: configured for UDMA/133
Jul  5 17:14:56 dcerouter kernel: [37318.132546] ata5: EH complete
Jul  5 17:16:02 dcerouter kernel: [37384.988920] ata5: hard resetting link
Jul  5 17:16:02 dcerouter kernel: [37384.988922] ata5: nv: skipping hardreset on occupied port
Jul  5 17:16:03 dcerouter kernel: [37385.456528] ata5: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
Jul  5 17:16:03 dcerouter kernel: [37385.472258] ata5.00: configured for UDMA/133
Jul  5 17:16:03 dcerouter kernel: [37385.472269] ata5.00: device reported invalid CHS sector 0
Jul  5 17:16:03 dcerouter kernel: [37385.472271] ata5.00: device reported invalid CHS sector 0
Jul  5 17:16:03 dcerouter kernel: [37385.472274] ata5.00: device reported invalid CHS sector 0
Jul  5 17:16:03 dcerouter kernel: [37385.472275] ata5.00: device reported invalid CHS sector 0
Jul  5 17:16:03 dcerouter kernel: [37385.472305] ata5: EH complete
Jul  5 17:20:56 dcerouter kernel: [37678.265759] ata5: hard resetting link
Jul  5 17:20:56 dcerouter kernel: [37678.265761] ata5: nv: skipping hardreset on occupied port
Jul  5 17:20:56 dcerouter kernel: [37678.732030] ata5: SATA link up 3.0 Gbps (SStatus 123 SControl 300)

After a while I find this:

Jul  5 23:09:07 dcerouter kernel: [58570.004250] ata5.00: configured for UDMA/133
Jul  5 23:09:07 dcerouter kernel: [58570.004268] ata5: EH complete
Jul  5 23:11:11 dcerouter kernel: [58693.988643] ata5: hard resetting link
Jul  5 23:11:11 dcerouter kernel: [58693.988645] ata5: nv: skipping hardreset on occupied port
Jul  5 23:11:12 dcerouter kernel: [58694.456031] ata5: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
Jul  5 23:11:12 dcerouter kernel: [58694.472235] ata5.00: configured for UDMA/133
Jul  5 23:11:12 dcerouter kernel: [58694.472258] ata5: EH complete
Jul  5 23:19:58 dcerouter kernel: [59220.536631] ata5: hard resetting link
Jul  5 23:19:58 dcerouter kernel: [59220.536633] ata5: nv: skipping hardreset on occupied port
Jul  5 23:19:58 dcerouter kernel: [59221.004530] ata5: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
Jul  5 23:19:58 dcerouter kernel: [59221.020742] ata5.00: configured for UDMA/133
Jul  5 23:19:58 dcerouter kernel: [59221.020755] ata5: EH complete
Jul  5 23:23:38 dcerouter kernel: [59440.956073] ata5: hard resetting link
Jul  5 23:23:38 dcerouter kernel: [59440.956075] ata5: nv: skipping hardreset on occupied port
Jul  5 23:23:39 dcerouter kernel: [59441.420110] ata5: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
Jul  5 23:23:39 dcerouter kernel: [59441.436254] ata5.00: configured for UDMA/133
Jul  5 23:23:39 dcerouter kernel: [59441.436268] ata5: EH complete
Jul  5 23:26:34 dcerouter kernel: [59616.740733] ata5: hard resetting link
Jul  5 23:26:34 dcerouter kernel: [59616.740735] ata5: nv: skipping hardreset on occupied port
Jul  5 23:26:35 dcerouter kernel: [59617.208033] ata5: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
Jul  5 23:26:35 dcerouter kernel: [59617.224228] ata5.00: configured for UDMA/133
Jul  5 23:26:35 dcerouter kernel: [59617.224245] ata5: EH complete
Jul  6 08:50:07 dcerouter kernel: imklog 4.2.0, log source = /proc/kmsg started.
Jul  6 08:50:07 dcerouter rsyslogd: [origin software="rsyslogd" swVersion="4.2.0" x-pid="1325" x-info="http://www.rsyslog.com"] (re)start
Jul  6 08:50:07 dcerouter rsyslogd: rsyslogd's groupid changed to 103

As you can see 23:26:35 was the last entry..and 8:50:07 is the time I restarted the system.
So far I checked the drives for smart errors, but there were none. I checked around google a bit and it seems it might be a software issue. So what do you guys think? The next thing to change would be the mainboard then (inkl. cpu).
Any input is welcome.

Cheers
Mario

[l3mce]

Jul  5 17:02:45 dcerouter kernel: [36587.564908] mdadm: sending ioctl 1261 to a partition!

Is this a raid? Did you try and setup a sw raid in web admin, or was one of the drives at some point part of a sw raid?

[maverick0815]

No, I'm not using any kind of raid system. When I started with linuxmce, back with 7.10, I had a hdd for the system and an IDE-Hardware Raid with 4 320gb hdds. Later I moved the data stored on that via webadmin to a 1TB disk. That was the only time I used RAiD.
Currently I have a 2TB disk with around 250GB for the system, 5GB swap and the rest for storage. By the way...remember when I went to the irc with the circular links? Well, I later realized my error- I had the remaining 1.8 TB mounted as /home...which of course was wrong, because then it was detected and I answered wrongly.
Anyways. I did a fresh install, wiped everything from that disk and here we are.
I am however a bit more inclined to think I have a hardware issue, because I just came home finding the system hanging again. I checked the logs..and so far I can see nothing suspicious around the time the thing crashed.
I saw this one

mdadm: sending ioctl 1261 to a partition!

5 minutes before the server freezes, but nothing else.
Right now I'm just guessing...

[Marie.O]

The mdadm message is a known (and afaik harmless) bug.

[l3mce]

just tossin it out there...

[maverick0815]

well...tossing out the mainboard I did.
call me master of desaster from now on...first the harddisk dies and as it turned out the mainboard's the culprit for it all. I found out for sure, when it didn't even want to boot from my usb-stick anymore.
So now I'm at my 10th install of linuxmce 1004..
at least everything is working as it should be :-)

[mkbrown69]

I've had errors like those, including the hard resetting link when one my raid 1 sets was doing it's weekly checks.  Occasionally, it would become totally unresponsive, requiring a power down of the system in order to restore it, and then I'd have to re-add it to the array and re-sync.  It's happened a few times (and only under heavy load), and I was close to RMA'ing the drive.  Then, this past Sunday morning I found my server off, and unable to power it back up.  The power supply had quit (and probably been on it's way out for a while due to capacitor aging), and was likely browning out under load.  So it's possible your power supply is on it's way out, or it's under-powered for the load.

There's a good power supply calculator at  http://extreme.outervision.com/psucalculatorlite.jsp

Try entering everything you have, and see what the recommended size is.  You can also add 30% for capacitor aging if your power supply is over a year old.  You'll be surprised how much it changes when you factor that in.

Hope that helps!

/Mike

[maverick0815]

thanks for the input, but I had exchanged the powersupply already and the problems didn't go away...now I exchanged the mobo and everything is fine so far