[How-To] Boot Entirely from SATA

[karog]

WarheadsSE, thanks for such a straightforward process to add a SATA disk. Worked first time for me. Like some others, I made a USB power cable and it works just fine as I am doing a direct connect without enclosure.

I wanted to prepare my disk on the pogo itself so I needed to attach the drive after boot. It took a bit of searching to find the command to rescan the SATA port:

Code: Select all

echo "0 0 0" > /sys/class/scsi_host/host0/scan

Perhaps you could add a script with this command called something like scansata to the oxnas_sata_boot.tgz download file. That would be really useful.

After preparation, I booted off a backup flash drive and used cp -a to copy my primary flash drive to the sata drive and it worked perfectly.

Also, I used the stage1.wrapped800 but I get BogoMIPS of 299.8 when I expected something like 320 at the 800 setting and in dmesg I see "Calibrating local timer... 375.50MHz." which also suggests a setting of 750. And I noticed on page 10 of this thread on Jan 17, 2012 3:08 pm the following message which identified the same observation:

Please guide me!!

How can I see the real processor speed?

I set stage1.wrapped -> stage1/stage1.wrapped800 and run disk_create
but seems cpu only 750 Mhz:

From serial console:

Code: Select all

Attempting to set PLLA to 750MHz ...                 


What is the problem?

Robert

and you replied in the next message

erwhoops.

Hmm.. seem's I'll have to update that

Did that by any chance fall through the cracks?

Again, thanks again for your excellent work.

[WarheadsSE]

I thought I had updated that, did you verify the checksum on the first page?

[karog]

I thought I had updated that, did you verify the checksum on the first page?

Yes, I verified the checksum that is currently on the first page and dated 12/31/11. And I just downloaded it within the last couple of days. Note also that the date of the previous 750 observation was Jan 17, 2012, after the date on the first page by the checksum.

I just verified again that the checksum is correct. And I untar'd the file to a new location and cmp'd the stage1.wrapped800 against the one I used yesterday and there was no difference. And I checked the link and it points to the 800.

I also cmp'd the 800 against the 750 to see if they might be the same and they are not.

Thanks for your timely response.

[Geoff]

I've been running mine 24/7 for a couple of months now at 850MHz without any hiccups ... just in case you feel like trying that.

[WarheadsSE]

Yes, the 2.6.31 kernel I have out there at the moment might be limited to a degree, but its very stable.

[karog]

I've been running mine 24/7 for a couple of months now at 850MHz without any hiccups ... just in case you feel like trying that.

Thanks for sharing your experience. I just tried 850 and it is up and running at 339.14 BogoMIPS as expected. I hope it will continue to run as stably as yours. If for any reason it does not, I will report that.

[WarheadsSE]

I've been running my 2.6.31 sata-booting system near 24/7, often at very high load for long periods, without any issues.

[karog]

After booting off SATA, I have been getting the following error shown via dmesg:

Code: Select all

[Thu May  3 23:38:29 2012] eth0: no IPv6 routers present
[Fri May  4 01:22:05 2012] ox820sata: reseting SATA core
[Fri May  4 01:22:05 2012] ata1.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen
[Fri May  4 01:22:05 2012] ata1.00: cmd 35/00:08:48:9e:2e/00:00:1e:00:00/e0 tag 0 dma 4096 out
[Fri May  4 01:22:05 2012]          res 40/00:01:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[Fri May  4 01:22:05 2012] ata1.00: status: { DRDY }
[Fri May  4 01:22:05 2012] ata1: hard resetting link
[Fri May  4 01:22:05 2012] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[Fri May  4 01:22:05 2012] ox820sata: reseting SATA core
[Fri May  4 01:22:11 2012] ata1.00: qc timeout (cmd 0xec)
[Fri May  4 01:22:11 2012] ox820sata: reseting SATA core
[Fri May  4 01:22:11 2012] ata1.00: failed to IDENTIFY (I/O error, err_mask=0x4)
[Fri May  4 01:22:11 2012] ata1.00: revalidation failed (errno=-5)
[Fri May  4 01:22:11 2012] ata1: hard resetting link
[Fri May  4 01:22:12 2012] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[Fri May  4 01:22:12 2012] ox820sata: reseting SATA core
[Fri May  4 01:22:22 2012] ata1.00: qc timeout (cmd 0xec)
[Fri May  4 01:22:22 2012] ox820sata: reseting SATA core
[Fri May  4 01:22:23 2012] ata1.00: failed to IDENTIFY (I/O error, err_mask=0x4)
[Fri May  4 01:22:23 2012] ata1.00: revalidation failed (errno=-5)
[Fri May  4 01:22:23 2012] ata1: limiting SATA link speed to 1.5 Gbps
[Fri May  4 01:22:23 2012] ata1: hard resetting link
[Fri May  4 01:22:23 2012] ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
[Fri May  4 01:22:23 2012] ata1.00: configured for UDMA/133
[Fri May  4 01:22:23 2012] ata1.00: device reported invalid CHS sector 0
[Fri May  4 01:22:24 2012] ox820sata: reseting SATA core
[Fri May  4 01:22:24 2012] ata1: EH complete


Note that the first line above is the last line of the normal boot and the error begins with the 2nd line.

I have tried this at all different frequencies from 850 down to 700 and the error appeared anywhere from 2 hours to 2 days. At 700 I thought I was good as it went 5 days without error but then just happened again. The machine has been sitting idle almost all of the time as I waited to see what would happen.

Googling the errors shows guesses all over the map from cables to disk controller to hard disk sectors with others contradicting each one.

Note also that at the end it resets at 1.5 Gbps and the disk is still usable with no apparent problem other than the above.

Does anybody here have experience with this problem and could shed more light? Thanks so much.

[WarheadsSE]

Considering that the message means that a command timed out, you're going to need to start at one end and work to the other.

smartctl -a /dev/sda

All the same, I've never seen that before.

[Geoff]

After booting off SATA, I have been getting the following error shown via dmesg:

Code: Select all

[Fri May  4 01:22:05 2012] ata1.00: status: { DRDY }


I have seen this kind of error on laptop drives at first occasionally, and then more and more often, when they're getting old and tired. I would suggest plugging the drive via eSATA into another computer that has "S.M.A.R.T." diagnostic tools and run a full self-test of the drive. E.g., Ubuntu has the "palimpsest" utility that appears in the menus as "Disk Utility." Note that these tools will not work when a drive is connected over USB; you need to connect it over eSATA.