Fatdog64-710 Alpha/2 [29 April 2016] [CLOSED]

[jake29]

Thanks jake29
If it happened once only then chalk it up to being a freak occurrence.
With 17 I have noticed the same freezing and it also happens sometimes when it should go into buffering mode . The video just catches up to the audio
Below is the link to Kodi 16.1 so now you can do comparisons

https://drive.google.com/open?id=0B0ggV ... HJhSWc0OG8

Thanks for this. CPU usage is significantly reduced when idle (0-1%) in Kodi 16.1 and when playing HD video (never going above 50%.) With Kodi 17.0, it would often idle at 90%.

[tonyfd710]

You're welcome.
Interesting stats. We will see when the beta comes out if they will have worked on that and the freezing.

[gcmartin]

Does TOP/HTOP show the % used by Kodi on the version tests you have done. Not asking about CPU, rather, about Kodi line-item of % and storage on your platform(s).

[jake29]

Does TOP/HTOP show the % used by Kodi on the version tests you have done. Not asking about CPU, rather, about Kodi line-item of % and storage on your platform(s).

Ah, yes htop seems more accurate. I can provide more detailed stats for Kodi 17.0 if necessary, but it was certainly (audibly - fan noise) pushing my system all the time it was running. It could be an ATI card specific issue however.

Kodi 16.1 idle: (2-5% CPU usage)


Kodi 16.1 playing 720p Youtube video: (10-14% CPU usage)

[step]

edit: read footnote

With 710 alpha 2 fsck con't complete checking a f2fs on an 8GB USB key.

Code: Select all

# fsck -r /dev/sdd2
fsck from util-linux 2.25.2
Info: sector size = 512
Info: total sectors = 15564800 (in 512bytes)
Info: MKFS version
  "Linux version 3.18.7 (root@fatdog64) (gcc version 4.8.3 (GCC) ) #1 SMP Tue Feb 17 20:32:52 EST 2015"
Info: FSCK version
  from "Linux version 3.18.7 (root@james-pc) (gcc version 4.8.3 (GCC) ) #1 SMP Mon Nov 16 01:58:11 WIB 2015"
    to "Linux version 4.4.3 (root@james-pc) (gcc version 4.8.3 (GCC) ) #2 SMP Tue Mar 1 23:25:01 AEDT 2016"
[ASSERT] (sanity_check_nid: 222) ret >= 0
/dev/sdd2: status 255, rss 2684, real 7.260277, user 0.379999, sys 0.253333
# uname -svor
Linux 4.4.3 #2 SMP Tue Mar 1 23:25:01 AEDT 2016 GNU/Linux

In another test I booted FD-702 but fsck stopped at the same point with a similar message, the only difference being that the 'from' line reported version 4.4.3 and the 'to' line reported version 3.18.7, and uname -a reported kernel version 3.18.7.

Footnote - update
It turns out that this cryptic fsck message is due to a f2fs partition error rather than a Fatdog64 error. I noticed the bad sector only when I looked at the kernel log file:

[ 3635.500348] blk_update_request: critical medium error, dev sdd, sector 15136656
[ 3635.500353] Buffer I/O error on dev sdd2, logical block 1883634, async page read

Reformatting the drive (luckily I had a backup) seems to have fixed it. Now fsck.f2fs reports

# /sbin/fsck.f2fs /dev/sdd2
Info: sector size = 512
Info: total sectors = 15564800 (in 512bytes)
Info: MKFS version
"Linux version 4.4.3 (root@james-pc) (gcc version 4.8.3 (GCC) ) #2 SMP Tue Mar 1 23:25:01 AEDT 2016"
Info: FSCK version
from "Linux version 4.4.3 (root@james-pc) (gcc version 4.8.3 (GCC) ) #2 SMP Tue Mar 1 23:25:01 AEDT 2016"
to "Linux version 4.4.3 (root@james-pc) (gcc version 4.8.3 (GCC) ) #2 SMP Tue Mar 1 23:25:01 AEDT 2016"

[FSCK] Unreachable nat entries [Ok..] [0x0]
[FSCK] SIT valid block bitmap checking [Ok..]
[FSCK] Hard link checking for regular file [Ok..] [0x0]
[FSCK] valid_block_count matching with CP [Ok..] [0x2]
[FSCK] valid_node_count matcing with CP (de lookup) [Ok..] [0x1]
[FSCK] valid_node_count matcing with CP (nat lookup) [Ok..] [0x1]
[FSCK] valid_inode_count matched with CP [Ok..] [0x1]
[FSCK] free segment_count matched with CP [Ok..] [0xeb3]
[FSCK] next block offset is free [Ok..]
[FSCK] fixing SIT types
[FSCK] other corrupted bugs [Ok..]

[jake29]

I don't think I had this happen with Fatdog 701, but in Fatdog 710a2 applications hog audio out. For example a paused audio / video file in VLC will prevent a wine app from outputting audio: Sound card is busy. Another example will be a browser tab with video / audio - even if playback has ended - will disable audio in KODI. I have to close the tab, close KODI and reopen KODI to have audio functioning.

From research, I've been lead to believe I can use the .asoundrc from Fatdog 701 as a workaround.

[step]

I don't think I had this happen with Fatdog 701, but in Fatdog 710a2 applications hog audio out.

I can't confirm. Right now I have VLC, mpv, a youtube browser tab, and two wine media players all playing audio at the same time. Pausing and restarting some of them leaves the others fully functional. Kernel 4.4.3, no ~/.asoundrc.

[jake29]

I don't think I had this happen with Fatdog 701, but in Fatdog 710a2 applications hog audio out.

I can't confirm. Right now I have VLC, mpv, a youtube browser tab, and two wine media players all playing audio at the same time. Pausing and restarting some of them leaves the others fully functional. Kernel 4.4.3, no ~/.asoundrc.

I thought the ~/.asoundrc file had been generated during install / first boot. I deleted it, rebooted and that appears to have resolved the issue.

Thanks for the reply.

[Belham]

I don't think I had this happen with Fatdog 701, but in Fatdog 710a2 applications hog audio out.

I can't confirm. Right now [b]I have VLC, mpv, a youtube browser tab, and two wine media players all playing audio at the same time[/b]. Pausing and restarting some of them leaves the others fully functional. Kernel 4.4.3, no ~/.asoundrc.

Step, you really gotta cutback on watching so many porn channels at the same time (haha)[/u]

[TeX Dog]

Code was not getting called due to lost paths after X was closed. Fix was to move code to when X still exists. and remove wrapper PID which is also broke, but not able to tell if it was related to first error.

[Satori]

Greetings,

FatDog 710A2 works well OTB, however when using Gslapt to do a full update of the packages:

1) it causes the lxqt-panel (taskbar) to disappear
2) when exiting chrome, chrome processes are still running, leaving the profile in a precarious state

switching to JWM (using the update)

1) lxqt-panel shows up but looks unconfigured
2) chrome exits cleanly (leaves no hanging processes)

switching back to OpenBox and a forced update to lxqt-panel application

1) lxqt-panel starts up with a blank panel
2) the lxqt-panel config is accessable, but missing items to edit

-------

without updating

1) lxqt-panel works correctly and it's config menu is accessable
2) chrome exits cleanly
3) when switching to JWM, it leaves the system with a blank screen with no option other than a manual power down

it seems chrome 51.0 does not play nice with versions of OpenBox, FatDog 702 also has this problem.

[step]

On my Intel Core 2 Duo notebook wlan is enabled but disconnected most of the time. After about 30 minutes after a fresh boot I notice 100% CPU load for process /usr/sbin/mdsnd, the Bonjour discovery service, I believe. Running "services mdns restart" fixes the problem, but it will happen again after some delay. Having to do the manual fix at random times is kind of annoying. Can anyone suggest an automated fix?

[kirk]

Step,

Haven't had a problem with mDNS, but you probably don't need it anyway. It's there mainly for printer discovery. The workaround would be to just disable it in the Service Manger.

Satori,

We'll be getting 710 beta out soon, avoid doing a full update for now, too many things in flux.

[jake29]

I'm unsure if this is a scenario I have created, but I've started getting this error whenever trying to shutdown, reboot or even open Geany.

Code: Select all

libpangoft2-1.0.so.0: undefined symbol: hb_buffer_set_cluster_level

Issues began when I started trying to get http://www.smokey01.com/fd700/sfs/chrom ... deVine.sfs to launch.

Rebooting does not appear to resolve the situation.

Edit: Correction - rebooting did resolve the issues. It seems that sfs is the source.

[stemsee]

I found this inconsistency not really a bug. Label is bluetooth on usb tethering page.

[smokey01]

I've just noticed a little bit of weirdness in Alpha2.

When I suspended the laptop from the menu it went to sleep as expected. To get it back I needed to touch the power button, this woke it up. What seemed strange was, the suspend script was still displaying in the task manager and using about 12% CPU. Normally mine idles at 0-1% when doing nothing with no apps loaded.

I killed the suspend script in task manager but the 12% load remained. Doing a restart of X dropped the CPU back to 0% but my WiFi stopped working and I couldn't get it going again with wpa_gui. Once I rebooted all was good again.

Closing the lid produced similar results however the CPU load was only 1%, not 12% as before. To get the WiFi working without rebooting I right-clicked on wpa_gui and clicked on restart connection, this worked. Maybe this needs to happen whenever the lid is reopened or after pressing the power button after suspend.

EDIT: I just did another test by closing the lid. Opened it after about 30 seconds and this time WiFi was working, Mmm not sure why the difference. It also appears to be working when clicking on the suspend Menu item.

Surely once the machine is woken from suspend the WiFi should automatically connect. Shouldn't the suspend script disappear from memory as well?

A bug or not a bug, you decide.

[kirk]

smokey01,

Unfortunately, suspend problems are kernel problems. Usually one of the modules have a problem. If you can find out which one is hanging up and it can be unloaded, you can rmmod it before suspend and then modprobe it after. You would have to edit /etc/acpi/actions/suspend.sh. Look for "echo mem > /sys/power/state " in that script, that's what actually causes the computer to suspend.
Fatdog 710 beta will have a newer kernel, so that may fix it for you,....or make it worse .

[smokey01]

Kirk it's intermittent which makes it even harder to diagnose.

[Belham]

smokey01,

Unfortunately, suspend problems are kernel problems. Usually one of the modules have a problem. If you can find out which one is hanging up and it can be unloaded, you can rmmod it before suspend and then modprobe it after. You would have to edit /etc/acpi/actions/suspend.sh. Look for "echo mem > /sys/power/state " in that script, that's what actually causes the computer to suspend.
Fatdog 710 beta will have a newer kernel, so that may fix it for you,....or make it worse .

Kirk & James,

What kind of time frame we looking at for the upcoming 'beta' release? I'm not trying to rush or anything like that, just wondering what a good guesstimate would be? October? December?? Next Spring?? I know it is crazy for you both, with the ARM / Android stuff too. Thanks!!

[smokey01]

Quite often when I reboot from FD710 to another disto, lets say slacko57, and then reboot back into FD710 it hangs while loading the xserver. I have had the same issue with earlier FD's and on various machines. When I power down and restart all is well again. Is something being trapped in memory?

Has anyone else experienced this?