Frequent freeze-ups of M2U

Tomas_Janovsky (Tomáš Janovský) 2017-04-22 18:45:43 UTC #37

Yes, probably nobody knows. I have built a new power adapter 5/12V 2x6A by my own. But I have still the same problem as you. Last ocurence I've cought this:

> Apr 22 05:08:18 localhost kernel: [24585.336241] ata1.00: exception Emask 0x10 SAct 0x1000 SErr 0x280100 action 0x6 frozen
> Apr 22 05:08:18 localhost kernel: [24585.351033] ata1.00: irq_stat 0x08000000, interface fatal error
> Apr 22 05:08:18 localhost kernel: [24585.363242] ata1: SError: { UnrecovData 10B8B BadCRC }
> Apr 22 05:08:18 localhost kernel: [24585.374620] ata1.00: failed command: READ FPDMA QUEUED
> Apr 22 05:08:18 localhost kernel: [24585.385909] ata1.00: cmd 60/00:60:62:39:56/01:00:cc:00:00/40 tag 12 ncq 131072 in
> Apr 22 05:08:18 localhost kernel: [24585.385909]          res 40/00:64:62:39:56/00:00:cc:00:00/40 Emask 0x10 (ATA bus error)
> Apr 22 05:08:18 localhost kernel: [24585.413978] ata1.00: status: { DRDY }
> Apr 22 05:08:18 localhost kernel: [24585.423449] ata1: hard resetting link
> Apr 22 05:08:19 localhost kernel: [24585.770168] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
> Apr 22 05:08:19 localhost kernel: [24585.771389] ata1.00: configured for UDMA/133
> Apr 22 05:08:19 localhost kernel: [24585.771482] ata1: EH complete
> Apr 22 08:03:48 localhost kernel: [35114.374649] Bad mode in data abort handler detected
> Apr 22 08:03:48 localhost kernel: [35114.374649] Internal error: Oops - bad mode: 0 [#1] SMP ARM
> Apr 22 08:03:48 localhost kernel: [35114.374649] Modules linked in: ip6table_filter ip6_tables iptable_filter ipt_MASQUERADE iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack ip_tables x_tables ov5640 option usb_wwan vfe_v4l2 vfe_io videobuf2_dma_contig ss videobuf2_memops videobuf2_core bcmdhd hci_uart
> Apr 22 08:03:48 localhost kernel: [35114.374649] CPU: 2 PID: 3328 Comm: mc Not tainted 3.10.104-BPI-M2U-Kernel #11
> Apr 22 08:03:48 localhost kernel: [35114.374649] task: df391ac0 ti: ddb00000 task.ti: ddb00000
> Apr 22 08:03:48 localhost kernel: [35114.374649] PC is at 0xffff10a4
> Apr 22 08:03:48 localhost kernel: [35114.374649] LR is at 0xb6b140c8
> Apr 22 08:03:48 localhost kernel: [35114.374649] pc : [<ffff10a4>]    lr : [<b6b140c8>]    psr: 20000197
> Apr 22 08:03:48 localhost kernel: [35114.374649] sp : ddb01fb0  ip : 00000000  fp : 00000071
> Apr 22 08:03:48 localhost kernel: [35114.374649] r10: 6472cac1  r9 : 0000d5c6  r8 : c7487000
> Apr 22 08:03:48 localhost kernel: [35114.374649] r7 : 00000001  r6 : c0000000  r5 : 0000d5c6  r4 : 90e00000
> Apr 22 08:03:48 localhost kernel: [35114.374649] r3 : 0001a9c7  r2 : 00000000  r1 : 60000000  r0 : 70e00000
> Apr 22 08:03:48 localhost kernel: [35114.374649] Flags: nzCv  IRQs off  FIQs on  Mode ABT_32  ISA ARM  Segment user
> Apr 22 08:03:48 localhost kernel: [35114.374649] Control: 10c53c7d  Table: 5d16c06a  DAC: 00000015
> etc...

Oleksandr_Kuzmenko (Олександр Кузьменко) 2017-04-22 19:09:02 UTC #38

Maybe my information can bee useful : all Ubuntu images on Google Drive is not stable. There are many freeze at compilation some libs, apps. Debian images are more stability. And the best is Centos image. I'm not got any freeze about 7 days.

Tomas_Janovsky (Tomáš Janovský) 2017-04-25 06:33:42 UTC #39

Hi all, I have changed governor to performance and since it seems to be better, can somebody also try it?

echo "performance" > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor

But I have still strange behavior, It freeze for some time but after several minutes the problems dissapear and BPI is running well again. No logs, no messages, nothing. Probably some ghosts in the kernel.

BPI M2U is really not good choose for stable server until stable version will be released.

avaf 2017-04-26 23:30:40 UTC #40

Please, give the complete log, use gist or pastebin, so we can know where/which module is crashing. Try to past the formatted output if possible. This will help to find a clue.

Tomas_Janovsky (Tomáš Janovský) 2017-04-27 13:03:56 UTC #41

It failed about 8 AM Apr 22
https://gist.github.com/e8ebf9448c4d8b6b1d0e8d331cf01c7b.git

avaf 2017-04-27 14:53:35 UTC #42

Can we see if there was something previously: grep "Apr 21" kern.log.1 and do you recall what possible process you were running?

Tomas_Janovsky (Tomáš Janovský) 2017-04-27 15:45:30 UTC #43

Nothing special:
https://gist.github.com/a603d08f60ed3e70153c2e560f461854.git

It failed during I was asleep and server did nothing.

enbarberis (Enrico) 2017-05-07 07:04:06 UTC #44

I'm in the same situation, my banana pi m2 ultra crash every 2 days. I'm using Debian 8 Jessie lite and with kernel 3.10.65. There is a stable version? I really need it since the banana pi is far away from and I can't reset it very often. From every log I can't infer anything. The only strange thing is that I have a lot of failed ssh connection (it seems a brute force attack). But I'm using fail2ban and public key so I think that there are no problems.

jata 2017-05-08 15:34:33 UTC #45

Hi all,

I have just purchased a banana pi m2 ultra. I am looking for a stable debian jessie lite install for this board. Where do I find them / any?

Is it likely that a new build for debian jessie will be released soon?

Tomas_Janovsky (Tomáš Janovský) 2017-05-12 13:10:27 UTC #46

You can try my workarround:
echo "performance" > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor

or modify /etc/default/cpufrequtils with:
GOVERNOR=performance

and then restart BPI.

Now I am uptime:
15:21:11 up 14 days, 23:34, 1 user, load average: 1,29, 1,33, 1,32

enbarberis (Enrico) 2017-05-16 19:02:44 UTC #47

I've done it and my banana pi crashes more often than before. I'm using 2017-01-20-debian-8-jessie-lite-beta3-bpi-m2u.img.zip that I found here https://drive.google.com/drive/folders/0B_YnvHgh2rwjR0JsaUltalFXanc. Now I've set again interactive as governor. Still waiting a more stable firmware. Thanks for the help

jata 2017-05-18 06:48:24 UTC #48

Any reason that you are not using the more recent jessie image? There is one released in April. It's in the same google drive folder as in your post.

I'm still waiting for my M2U to be delivered from China. Reading this thread, i am not pleased that I bought it. Seems like the hardware has some issues and firmware is unstable.

The support/help from the BPI guys is terrible but overall the community is helpful.

Oleksandr_Kuzmenko (Олександр Кузьменко) 2017-05-18 15:47:50 UTC #49

I'm use Centos image. And uptime more 2 month. All is fine.

facat (facat) 2017-05-19 00:57:23 UTC #50

Where is the CentOS image for M2U?

Customer (customer) 2017-05-19 02:37:03 UTC #51

Images here: Obviously maintained by someone other than webmaster. Old links at the top of the page. Occasionally a BananaPi person will drop in to answer some little something, but mostly the BananaPi posts are just BPi related things some barely English-speaking person scrapes from the web.

Oleksandr_Kuzmenko (Олександр Кузьменко) 2017-05-19 05:13:10 UTC #52

I'm use kernel 3.10.104

facat (facat) 2017-05-19 06:03:39 UTC #53

Thanks for your reply. I cannot find anything related to CentOS in that page. I event searched Centos but found nothing.

Customer (customer) 2017-05-19 13:34:37 UTC #54

centos-lite. it's on the page i linked but here's a direct link

facat (facat) 2017-05-19 13:52:43 UTC #55

Got it. Thanks very much.

Tomas_Janovsky (Tomáš Janovský) 2017-07-13 07:49:49 UTC #56

Since I'v modified governor to performance it is stable. No other modifications, only power source from notebook and 6/12V stabilizator for BPI/3.5" disk:

root@bananapi:~# uptime
10:05:45 up 76 days, 18:19, 1 user, load average: 1,10, 1,11, 1,13