After activated RTSP/TCP the satip box freeze every day! #43
Comments
|
look at #34 if this is similar |
|
If it's software (kernel in this case) fault, the serial console might show something. Otherwise, it's very difficult to determine the fault like this. I have my two boxes open with fan and all major chips (CPU/network/demodulators/tuners) have heatsinks, too. But saying that, I won't force anyone to break the warranty. http://postimg.org/image/aaya8pnb5/ - circles marks added / modified heatsinks. |
|
Thanks for the quick responses. Yesterday I connect the serial cable over a usb adapter to a server.
Is this output normal or is it needed to configure something? |
|
Hi same problem here (same inverto box) for me it was necessary to monitoring my satip-device with a simple fping and zabbix to get an info-mail on crash (see up/down-times on zabbix-picture) i dont think it is a temperature problem. my satip-device is installed near my multiswitch (spaun) with short cables (0,5m) on the loft (see temperature-picture) i also supporting a friend with the same configuration. his satip-box crashes as well (same inverto box) making some stresstests does not crashes the box (mux-scans, recording four full-hd streams, trigger eit-epg) but in my opinion, satip-box is more stable after cleaning not necessary mux, services and channels from tvheadend. after cleaning, my box was 7 days up-and-running (but yesterday it was crashing again) after reading the post from Imerwin, i bought a "USB zu TTL-converter-Modul" to get a serial output from the satip box (but i has to install it) Using: |
Configure your serial device properly. See the top-level README - https://github.com/perexg/satip-axe/blob/master/README.md |
If the bug is in the Inverto drivers, it would be very very difficult to fix this. But a captured crash from the serial console might help to identify the lock issue. |
|
Hi, I can confirm this issue. Running my Digibit for weeks without any issue with satip-axe I've changed yesterday to RTSP/TCP to get rid of some continuity errors (which still appear even all is running on a switched gigabit network with high quality D-Link managed switches). I had a complete crash 5 minutes ago, the box was not even reacting on a ping. Switched back to UDP for now ... |
|
Hi,
USB stick is (upper) connected with "satip-axe-201508061051-10.fw" and "satip-axe-201508061051-10.usb" renamed to idl4k.scr It is possible to start the debug mode from the flash memory (without usb boot)? Thanks! |
|
@Imerwin : Don't do anything with u-boot (don't press anything). Just keep it running - the box will boot to satip-axe as before, but you can see all kernel outputs. The 'debugfw' command set is from Inverto guys - I don't use it for the satip-axe firmware. |
|
Hi,
After the crash the network is unreachable but the serial console work. Ping to himself (127.0.0.1 and 10.10.10.115) work but to the gateway (10.10.10.1) not. Here is the log file with what i have tried. |
|
Hi, if it is a problem with the network card/driver, maybe we can getting more infos modifying driver parameters on kernel build: my box is running at this time. now iam also watching some network parameters/statistics: |
|
I agree, it looks like a network driver/hw issue. You may try 'ifconfig eth0 down ; ifconfig eth0 up' sequence to restart the network driver. Also, you may try to force 100M speed through ethtool (just for tests or you may put this settings to the shell script /etc/sysconfig/localboot). I'll look if I can backport some ethernet driver changes from the stlinux kernel. |
|
Hi,
If I manually disconnect the network cable and reconnect the cable then comes this message:
ifconfig eth0 down and then up also don't helps. When I change the network speed i see this messages in the dmesg output:
I have now changed the network speed to 100M let's see how long the box work. http://imerwin.com/tmp/satip/satip_serial_output_11.02.2016.log |
|
Hi, is there any possibility to compare limits/settings between the original firmware and satip-axe? i rebooted my box and changed some parameters to see, whats happend: let speed to 1Gbit/s |
|
@satip: You should probably do interface up / down sequence when you change this - these parameters are usually used only when the network interface is initialized. |
|
thx ifconfig eth0 down && ifconfig eth0 up checking parameters again: |
|
Firmware with the updated stmmac driver from the latest stlinux kernel: Please, give me feedback. Note: The updated driver has /sys/module/stmmac/parameters/wol_plus_en file (it does not exist in the old driver). Note2: There's debugfs support: mount -t debugfs debufs /sys/kernel/debug cat /sys/kernel/debug/stmmaceth/descriptors_status |
|
I did some more hacking and here's new package (beta for 11 release): http://s000.tinyupload.com/?file_id=78443875428972064634 List of changes: https://github.com/perexg/satip-axe/blob/master/dist/README (upcoming) Please, give me feedback. I enabled Realtek PHY initialization code in this kernel too (it won't affect probably anything, but who knows). |
|
Unfortunately, the updated driver won't solve the locking issue. Trying with setup from @satip (added to /etc/sysconfig/config file - echo lines): # # IPV4 config # #IPV4="dhcp" # or IPV4="192.168.1.200" IPV4_NETMASK="255.255.255.0" IPV4_BROADCAST="192.168.100.255" IPV4_GATEWAY="192.168.100.1" echo "512" > /sys/module/stmmac/parameters/dma_rxsize echo "512" > /sys/module/stmmac/parameters/dma_txsize echo "8192" > /sys/module/stmmac/parameters/buf_sz |
|
Another lockup here... Trying: ethtool -K eth0 sg off |
|
some informations from my side friends box (SATIP-AXE server 201602121045-10): my box (SATIP-AXE server 201508061051-10): attached serial console on both boxes. running "picocom" in a linux "screen" session. |
|
Hi, |
|
My testing: With scatter-gather off (ethtool -K eth0 sg off) I wasn't able to lock the ethernet with new fw in two days (did two stress tests and a standard operation). I will make 'sg off' as default in the driver and also I increase the DMA descriptors sizes to 512. |
|
For more recent pre-releases see issue #45 . The scatter-gather is off by default in the kernel. |
|
Even If it is not directly connected to this issue here, with the new official fw 1.24.0.156 I get continuity errors like if I disable TCP at axe (Raspberry Pi 2). Seems like they changed something. Semi-official V1.0.0.146 works like intended. |
|
hi, for me, you can close this bug friends box: 5 days uptime with new firmware (201602121045-10) THX for support |
|
Ok, closing - resolved in build 11 (issue #45). |
Hi,
I have a problem with the Tvheadend RTSP/TCP function.
If I activate this function my satip box crashes minimum one time per day.
When the box crashes, the network is unreachable but the led with the last tuner state is shining.
If I use RTSP with UDP then I can watch 2 HD Streams without a problem after I start a third stream I get a lot of continuity errors but the box don’t crash.
With RTSP/TCP I can watch 4 HD Streams or more without continuity errors.
I use:
HTS Tvheadend 4.1-1242~ge14c134
Inverto IDL-400s with satip-axe-201508061051-10
my summary:
own VLAN only for satip and tvheadend or a new network cable – same problem
I cannot exactly describe when the box goes down.
I can watch 3 or 4 HD streams over hours with no problem but it’s also possible to watch 1 stream and the box crashes between 2 and 6 hours.
Maybe someone can help me.
Thanks!
The text was updated successfully, but these errors were encountered: