[Ovmsdev] webserver (?) crash
Mark Webb-Johnson
mark at webb-johnson.net
Wed Mar 21 08:34:24 HKT 2018
With 3.1, here is what I get on a clean boot:
OVMS# module memory
Free 8-bit 133180/282048, 32-bit 8164/24420, SPIRAM 4148396/4194252
And, again with 3.1, here it is after starting SIMCOM modem, wifi in APCLIENT mode, TR vehicle module, and Server v2:
OVMS# module memory
Free 8-bit 79520/282048, 32-bit 8164/24420, SPIRAM 4136036/4194252
OVMS# module tasks
Number of Tasks = 18 Stack: Now Max Total Heap 32-bit SPIRAM
Task 3FFAEB34 1 esp_timer 392 648 6656 55140 644 0
Task 3FFAF7E0 2 eventTask 444 3116 4608 7620 0 0
Task 3FFC3F68 3 CanRxTask 420 420 4096 0 0 0
Task 3FFC8060 4 ipc0 396 444 1024 10848 0 0
Task 3FFC8660 5 ipc1 396 444 1024 12 0 0
Task 3FFCA488 8 IDLE 356 484 1024 0 0 0
Task 3FFCAA1C 9 IDLE 360 488 1024 0 0 0
Task 3FFCC3B0 10 Tmr Svc 396 3276 6144 7816 0 0
Task 3FFCF8F0 14 Housekeeping 364 2716 6144 53420 0 0
Task 3FFC8AB0 16 tiT 496 2752 6144 7112 0 0
Task 3FFD79C8 17 SIMCOMTask 452 2260 4096 4404 0 0
Task 3FFDB8C8 18 AsyncConsole 756 4964 5120 45864 15488 0
Task 3FFDCD84 19 mdns 416 1736 4096 108 0 0
Task 3FFE147C 20 wifi 424 2296 4096 5180 0 0
Task 3FFE3D74 21 pmT 416 464 2560 0 0 0
Task 3FFE6C30 22 NetManTask 732 2556 7168 2488 0 0
Task 3FFEAB08 23 Vrx Task 452 452 4096 0 0 0
Task 3FFEAD0C 24 OBDII ECU Task 424 424 6144 0 0 0
I think something got broken with a recent build of ESP-IDF so that our ‘module memory’ and ‘module tasks’ is not showing SPIRAM usage correctly. But, it is definitely used and certainly helps ease the stress:
OVMS# module memory
Free 8-bit 133032/282048, 32-bit 8164/24420, SPIRAM 4148396/4194252
OVMS# obdii ecu start can3
OBDII ECU has been started
OVMS# module memory
Free 8-bit 125000/282048, 32-bit 8164/24420, SPIRAM 4147772/4194252
--Task-- Total DRAM D/IRAM IRAM SPIRAM +/- DRAM D/IRAM IRAM SPIRAM
AsyncConsole 13896 6144 15488 0 +1876 +6144 +0 +0
We are far from perfect, but I think good enough for some factory firmware. Anything else will come as OTA.
Regards, Mark.
> On 21 Mar 2018, at 3:12 AM, Michael Balzer <dexter at expeedo.de> wrote:
>
> Am 20.03.2018 um 06:47 schrieb Greg D.:
>> Ok, interesting. Am I seeing an out-of-memory crash? I can't tell...
>>
>> It seems like the webserver status button is the worst offender,
>> probably because it's doing a lot of work and building a big payload.
>> Very often hangs for a while before responding, and if so, usually
>> results in the job queue overflow. If I catch it quick enough and hit
>> home or some other shorter task, the overflow will drain and I'll get
>> the new screen; otherwise a crash. Sometimes there will be a timeout
>> mixed in. I'm wondering if our little squirrel in the ESP32 just can't
>> pedal hard enough...
>
> Most crashes now are related to lack of free memory.
>
> Regarding the status page I just hit the page 10 times:
>
> OVMS# mo m
> Free 8-bit 27264/283200, 32-bit 27544/55820, SPIRAM 0/0
> I (784440) webserver: HTTP GET /status
> D (785010) webserver: Serve /status: 6068 bytes used, 21200 free
> I (787240) webserver: HTTP GET /status
> D (787830) webserver: Serve /status: 6064 bytes used, 21168 free
> I (788720) webserver: HTTP GET /status
> D (789290) webserver: Serve /status: 6104 bytes used, 21164 free
> I (790110) webserver: HTTP GET /status
> D (790700) webserver: Serve /status: 6100 bytes used, 21164 free
> I (791530) webserver: HTTP GET /status
> D (792120) webserver: Serve /status: 6068 bytes used, 21200 free
> I (792980) webserver: HTTP GET /status
> D (793550) webserver: Serve /status: 6100 bytes used, 21164 free
> I (794400) webserver: HTTP GET /status
> D (794980) webserver: Serve /status: 6068 bytes used, 21200 free
> I (795860) webserver: HTTP GET /status
> D (796730) webserver: Serve /status: 6100 bytes used, 21164 free
> I (797160) webserver: HTTP GET /status
> D (797740) webserver: Serve /status: 6064 bytes used, 21168 free
> I (798430) webserver: HTTP GET /status
> D (799020) webserver: Serve /status: 6060 bytes used, 21200 free
> OVMS# mo m
> Free 8-bit 27264/283200, 32-bit 27544/55820, SPIRAM 0/0
>
> So no memory loss and no crashes.
>
> The 6K is just the buffer usage after processing. The commands executed for the status page need additional RAM, but above 20 K free, the page normally has no
> problems.
>
> Regards,
> Michael
>
> --
> Michael Balzer * Helkenberger Weg 9 * D-58256 Ennepetal
> Fon 02333 / 833 5735 * Handy 0176 / 206 989 26
>
>
> _______________________________________________
> OvmsDev mailing list
> OvmsDev at lists.teslaclub.hk
> http://lists.teslaclub.hk/mailman/listinfo/ovmsdev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.teslaclub.hk/pipermail/ovmsdev/attachments/20180321/1668a980/attachment.html>
More information about the OvmsDev
mailing list