| 22:05 | drgrant | Hi all. Anyone using their CuBox as a NFS/samba NAS? |
| 22:06 | drgrant | I am seeing lackluster performance on my dual-core, transfer rates never top 20 MB/s. |
| 22:08 | wbx | drgrant: more should be possible. are you using a gigabit switch? |
| 22:09 | drgrant | wbx: yeah, 470 Mbps on iperf too. |
| 22:10 | drgrant | I saw 50 MB/s over NFS once, but have not been able to reproduce it since. :S |
| 22:10 | fritsch | the question rather is |
| 22:10 | wbx | hmm. which kernel? |
| 22:10 | fritsch | where are you transfering to? |
| 22:10 | fritsch | to internal sdcard? |
| 22:10 | fritsch | to external connect esata? |
| 22:10 | drgrant | No, eSATA HDD. |
| 22:10 | fritsch | okay, test the eSata alone first |
| 22:10 | drgrant | 100 MB/s read, 80 MB/s write. |
| 22:10 | fritsch | you tested that form cubox? |
| 22:10 | drgrant | yup, hdparm and dd. |
| 22:11 | fritsch | with sync? |
| 22:11 | fritsch | hdparm is a bit academically |
| 22:11 | drgrant | hmm, no |
| 22:11 | fritsch | is there a difference between read and write? |
| 22:11 | drgrant | wbx: 3.10.42-geb80c2b, built it last night. |
| 22:11 | fritsch | let me make a short test |
| 22:11 | wbx | drgrant: you should better try 3.14.x for best performance. |
| 22:12 | drgrant | I used dd if=/dev/zero of=/test oflag=nocache bs=1M count=100 |
| 22:12 | drgrant | wbx: No difference, and 3.14.14+ actually gives me random ATA reset errors on the eSATA drive. |
| 22:13 | drgrant | As seen here: http://www.solid-run.com/community/topic1193.html |
| 22:15 | wbx | drgrant: then may be a 3.17.x kernel with rmk-patches might be worth a try |
| 22:15 | drgrant | hmm, where can I find the sources? |
| 22:15 | drgrant | and what is rmk? noob here. :P |
| 22:16 | wbx | drgrant: rmk is the linux arm maintainer, who is working on bringing all cubox stuff into linus tree. |
| 22:16 | wbx | http://www.home.arm.linux.org.uk/~rmk/cubox/ |
| 22:16 | drgrant | Ok, cool. |
| 22:17 | wbx | there are always performance tunings for fec (ethernet driver) and other stuff in there |
| 22:17 | drgrant | I see that 3.18 is up too. So you would recommend 3.17rc5? |
| 22:17 | wbx | may be his 3.18 patch set and 3.18 |
| 22:18 | drgrant | I guess I'll build and report back. Thanks. |
| 22:19 | wbx | my cubox-i is running with 3.14, but I haven't used samba or nfs, yet. |
| 22:19 | drgrant | Do you use eSATA? |
| 22:20 | wbx | i have a disk. but need to sort some things out. at the moment a usb disk is connected |
| 22:20 | drgrant | ah |
| 22:20 | wbx | need to find the power supply for the external sata drive ;) |
| 22:20 | drgrant | I use a powered enclosure. :) |
| 22:20 | wbx | plan is to buy a big disk with esata to make a nas |
| 22:21 | drgrant | watch your syslogs, you might be affected by the ATA errors too. |
| 22:21 | wbx | my 300 gb usb disk is to old and to small |
| 22:22 | wbx | at the moment I fighting with monitor blanking while using kodi :( |
| 22:28 | wbx | connected my 150 gb disk via esata. let's see |
| 22:31 | fritsch | wbx: does it happen every minute ( the blanking?) or when you bend the hdmi cable? |
| 22:34 | wbx | fritsch: interesting. the blanking does not happen when playing a local video from esata drive. only when trying to play a video via mediathek (ard / tatort) |
| 22:34 | fritsch | wbx: something in kodi.log? kernel does something? |
| 22:34 | wbx | fritsch: hdmi is connected via a dvi adapter to a eizo flexscan. |
| 22:35 | fritsch | does the kernel register some change? |
| 22:35 | fritsch | or is it purely a weak contact or something? |
| 22:39 | wbx | as star wars from local disk works fine, I don't think it is a cable problem |
| 22:39 | fritsch | does it switch refreshrate? |
| 22:39 | fritsch | for the mediathek one? |
| 22:39 | fritsch | e.g. to 50hz? |
| 22:40 | fritsch | does star wars also flicker if you watch it in != 24p mode? (if you use adjust refreshrate to match video at all) |
| 22:41 | wbx | hm.. may be just a network issue. now mediathek works fine since 3 minutes |
| 22:42 | wbx | i tried at 21 o'clock while it was running on tv. may be now the stream is more stable. |
| 22:44 | fritsch | should not happen |
| 22:44 | fritsch | do you see some "could not get any data from paket"? |
| 22:44 | fritsch | in the logs? |
| 22:44 | fritsch | we have no code for blanking just4fun |
| 22:44 | fritsch | :p |
| 22:44 | fritsch | which kodi version that is, btw.? |
| 22:45 | wbx | release version. |
| 22:46 | wbx | yes, I got a lot of could not get any data from paket |
| 22:46 | fritsch | that means network is stuck |
| 22:46 | fritsch | or no data coming trhough |
| 22:46 | fritsch | if you normally use iptv simple |
| 22:46 | fritsch | you need another patch |
| 22:46 | fritsch | we did not analyze the stream enough and that caused issues |
| 22:46 | wbx | so, yes. it seems to be a network issue. |
| 22:47 | fritsch | https://github.com/xbmc/xbmc/pull/6042 |
| 22:47 | fritsch | if you find some issues with specific streams |
| 22:47 | fritsch | this one you need |
| 22:47 | fritsch | also 14.1-RC1 is constantly filling the Helix branch |
| 22:48 | fritsch | so 14.1 will happen soon |
| 22:48 | wbx | hmm. and now after 15 minutes, video is black, sound is playing. |
| 22:48 | fritsch | if you press enter - you see the controls? |
| 22:49 | fritsch | or complete screen is gone? |
| 22:49 | wbx | complete screen is gone. display says no signal |
| 22:49 | fritsch | wbx: which distro is that? Anythin special patched in? kernel wise? |
| 22:50 | wbx | 3.14.26 (self compiled distro). only solidrun patches. |
| 22:51 | wbx | In DVI Mode disable interrupts |
| 22:51 | wbx | do you thing it is an kernel issue? |
| 22:52 | fritsch | i did not write any code to kodi lately that blanks the screen :-) |
| 22:52 | fritsch | do you have an kodi.log for completeness? |
| 22:52 | fritsch | does the kernel say something? |
| 22:54 | wbx | dmesg says only In DVI Mode disable interrupts |
| 22:54 | fritsch | and kodi.log? |
| 22:54 | fritsch | does SystemEGL report something? |
| 22:54 | wbx | http://openadk.org/kodi.log |
| 22:55 | fritsch | mmh |
| 22:55 | fritsch | it really seems to go offline |
| 22:55 | fritsch | you see that with AESinkALSA |
| 22:55 | fritsch | we call AddPackets |
| 22:55 | fritsch | and on failure, we call pcm_recover |
| 22:55 | fritsch | when that happens twice the audio out is gone |
| 22:55 | fritsch | your DVI does carry audio? |
| 22:56 | wbx | no. |
| 22:56 | fritsch | which audio output do you have configured? |
| 22:56 | fritsch | spdif? |
| 22:57 | wbx | audio is via a spdif optic dac to normal pc box |
| 22:57 | fritsch | wondering |
| 22:57 | fritsch | wh the alsa device goes nuts |
| 22:57 | wbx | i think the audio messages occured whily i tried to reblank via some keys |
| 22:57 | fritsch | not sure why |
| 22:57 | wbx | right cursor or sth. |
| 22:58 | fritsch | the click noise? |
| 22:58 | wbx | it does not occur while the creen is blanking |
| 22:58 | fritsch | oki |
| 22:58 | fritsch | still interested why adding some packets to a working sound device causes the device to not eat a package |
| 22:58 | wbx | it is not related. I could generate a new log, without trying to awake the screen via keyboad |
| 22:58 | fritsch | oki, i need to go |
| 22:58 | wbx | me too. ;) |
| 22:58 | wbx | sleeping time |
| 22:59 | fritsch | debuglog will help more I think |
| 22:59 | wbx | need to recompile tomorrow |
| 22:59 | fritsch | but that alsa stuff is also relevant, as the device seems to be gone |
| 23:24 | drgrant | Hmm, my new 3.18.1 kernel isn't booting. |