[weewx-user] Re: Please can I have my bug back?

2016-10-30 Thread mwall 
On Sunday, October 30, 2016 at 4:47:57 AM UTC-4, Macha wrote:
>
> In fact, the original "lockouts" weren't coincident with weewx's reporting 
> and seemed pretty random (and infrequent). I still suspect that the whole 
> issue had something to do with the fact that this "original" machine stores 
> "/" on my NFS and consequently Ethernet traffic is high.
>

all evidence (from 5+ years of dealing with these stations) points to a 
hardware problem with the fine offset hardware and/or firmware design, not 
a software problem on the computer.

the usb comm failures you were experiencing are not due to the cpu load of 
the pi or weewx's ability to process - they are due to the 
voltage/current/power on the usb.

m

-- 
You received this message because you are subscribed to the Google Groups 
"weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to weewx-user+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[weewx-user] Re: Please can I have my bug back?

2016-10-30 Thread Macha 
No I didn't. In fact I didn't play around with nice levels etc.

The "original" RPi is a Pi 2 (more by accident than intention, but it 
should have enough "welly") - it now runs just weewx and a single camera 
and its CPU utilisation is very low. 

In fact, the original "lockouts" weren't coincident with weewx's reporting 
and seemed pretty random (and infrequent). I still suspect that the whole 
issue had something to do with the fact that this "original" machine stores 
"/" on my NFS and consequently Ethernet traffic is high.

Cheers

PS: The RPi now running 3 cameras, several Adafruit sensors and audio 
streaming has had a traumatic history. It is a "Model B+ and was originally 
housed outdoors in a weatherproof clear plastic case. Unfortunately, this 
case wasn't as rugged as hoped and about 12 months ago - after it had 
crashed and following a few days of high wind and heavy rain - I found the 
RPi lying in the grass, soaking wet and covered in a kind of powdery grey 
gunk. I cleaned it up (toothbrush + alcohol + compressed air) and it then 
worked fine agin. The only lasting damage being the failure of the 
spring/clip thing that holds the SD card in place - this is now secured by 
a plastic cable tie. The original SD card was completely fried.

On Sunday, October 30, 2016 at 8:53:44 AM UTC+1, Andrew Milner wrote:
>
> when you had cams and weewx on one machine did you have turbo mode 
> enabled?  It would have needed it I expect.
>
> weewx will have failed because the camera were hogging resources - you 
> could have tried raising the priority.  weewx will trundle along at very 
> low cpu usage UNTIL it generates reports and plots - then it will need very 
> high resources.
>
>
>
> On Sunday, 30 October 2016 09:42:12 UTC+2, Macha wrote:
>
>> OK, status report on my "lockout" with weewx and my FOUSB station.
>>
>> I moved 2 webcams from the RPi running weewx onto a different machine 
>> (another RPi, I have quite a few).
>>
>> Result? No weewx lockouts for nearly 10 days. I'm not sure if this passes 
>> the "5 sigma" test, but it's certainly a record uptime.
>>
>> Still a few puzzlements - on the previous setup, it was always weewx 
>> (well Python I suppose) that "failed", the webcams (run with mjpg-streamer) 
>> were always rock-solid. On the new setup, the RPi which now has the "moved" 
>> webcams is quite busy - it runs the 2 new cameras, plus an RPi camera (all 
>> driven with mjpg-streamer) and also icecast2+darkice (monitoring bird 
>> tweets, etc.) - CPU utilisation is around 35%.  This machine never fails.
>>
>> The previous RPi is now chuffing along at <1% 
>>
>> Anyway, thanks again for the advice and comments ...
>>
>>

-- 
You received this message because you are subscribed to the Google Groups 
"weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to weewx-user+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[weewx-user] Re: Please can I have my bug back?

2016-10-30 Thread Andrew Milner 
when you had cams and weewx on one machine did you have turbo mode 
enabled?  It would have needed it I expect.

weewx will have failed because the camera were hogging resources - you 
could have tried raising the priority.  weewx will trundle along at very 
low cpu usage UNTIL it generates reports and plots - then it will need very 
high resources.



On Sunday, 30 October 2016 09:42:12 UTC+2, Macha wrote:

> OK, status report on my "lockout" with weewx and my FOUSB station.
>
> I moved 2 webcams from the RPi running weewx onto a different machine 
> (another RPi, I have quite a few).
>
> Result? No weewx lockouts for nearly 10 days. I'm not sure if this passes 
> the "5 sigma" test, but it's certainly a record uptime.
>
> Still a few puzzlements - on the previous setup, it was always weewx (well 
> Python I suppose) that "failed", the webcams (run with mjpg-streamer) were 
> always rock-solid. On the new setup, the RPi which now has the "moved" 
> webcams is quite busy - it runs the 2 new cameras, plus an RPi camera (all 
> driven with mjpg-streamer) and also icecast2+darkice (monitoring bird 
> tweets, etc.) - CPU utilisation is around 35%.  This machine never fails.
>
> The previous RPi is now chuffing along at <1% 
>
> Anyway, thanks again for the advice and comments ...
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to weewx-user+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[weewx-user] Re: Please can I have my bug back?

2016-10-30 Thread Macha 
OK, status report on my "lockout" with weewx and my FOUSB station.

I moved 2 webcams from the RPi running weewx onto a different machine (another 
RPi, I have quite a few).

Result? No weewx lockouts for nearly 10 days. I'm not sure if this passes the 
"5 sigma" test, but it's certainly a record uptime.

Still a few puzzlements - on the previous setup, it was always weewx (well 
Python I suppose) that "failed", the webcams (run with mjpg-streamer) were 
always rock-solid. On the new setup, the RPi which now has the "moved" webcams 
is quite busy - it runs the 2 new cameras, plus an RPi camera (all driven with 
mjpg-streamer) and also icecast2+darkice (monitoring bird tweets, etc.) - CPU 
utilisation is around 35%.  This machine never fails.

The previous RPi is now chuffing along at <1% 

Anyway, thanks again for the advice and comments ...

-- 
You received this message because you are subscribed to the Google Groups 
"weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to weewx-user+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[weewx-user] Re: Please can I have my bug back?

2016-10-20 Thread vince 
On Thursday, October 20, 2016 at 4:11:01 AM UTC-7, mwall wrote:

> test 1: remove all of the cameras and let the system run for a week
>
> test 2: put a powered usb hub between the weather station console and the 
> pi and let the system run for a week
>

 
Three usb cameras is a huge number.   Add another test:
 - put the cameras on a powered usb-hub

-- 
You received this message because you are subscribed to the Google Groups 
"weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to weewx-user+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[weewx-user] Re: Please can I have my bug back?

2016-10-20 Thread Andrew Milner 
One further thing would be interesting to see if it had any effect - to 
downgrade the priority of the three cameras by reniceing them to say 10 
or maybe renice weewxd to -15 to give weewx higher priority.  Not sure if 
it is weewx which needs the higher priority or something else - and anyway 
matthew's suggestion of removing the cameras completely would yield the 
same verdict as to the cause.

I am a firm bel;iever in powered hubs and good power supplies as saviours 
when it comes to the RPi.


On Thursday, 20 October 2016 14:11:01 UTC+3, mwall wrote:

> On Thursday, October 20, 2016 at 6:18:34 AM UTC-4, Macha wrote:
>>
>> Yes, the console has batteries, recently changed, so still pretty strong. 
>> Rebooting the pi has no effect on the console. it just sits there and 
>> doesn't undergo any kind of reset/reboot/power cycle.
>>
>
> you are not experiencing the dreaded fine offset firmware lockups.  your 
> comm failures are almost certainly due to voltage/current problems on the 
> pi.
>
> please try the following tests:
>
> test 0: next time you lose comms to the console, do not reboot the pi, and 
> do not restart weewx.  just unplug the console from the pi then plug it 
> back in.  comms should return, and weewx should recover.
>
> test 1: remove all of the cameras and let the system run for a week
>
> test 2: put a powered usb hub between the weather station console and the 
> pi and let the system run for a week
>
> i'll bet the comm problems go away when you remove the cameras or add the 
> powered hub
>
> m 
>

-- 
You received this message because you are subscribed to the Google Groups 
"weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to weewx-user+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[weewx-user] Re: Please can I have my bug back?

2016-10-20 Thread mwall 
On Thursday, October 20, 2016 at 6:18:34 AM UTC-4, Macha wrote:
>
> Yes, the console has batteries, recently changed, so still pretty strong. 
> Rebooting the pi has no effect on the console. it just sits there and 
> doesn't undergo any kind of reset/reboot/power cycle.
>

you are not experiencing the dreaded fine offset firmware lockups.  your 
comm failures are almost certainly due to voltage/current problems on the 
pi.

please try the following tests:

test 0: next time you lose comms to the console, do not reboot the pi, and 
do not restart weewx.  just unplug the console from the pi then plug it 
back in.  comms should return, and weewx should recover.

test 1: remove all of the cameras and let the system run for a week

test 2: put a powered usb hub between the weather station console and the 
pi and let the system run for a week

i'll bet the comm problems go away when you remove the cameras or add the 
powered hub

m 

-- 
You received this message because you are subscribed to the Google Groups 
"weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to weewx-user+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[weewx-user] Re: Please can I have my bug back?

2016-10-20 Thread Macha 
Hi,

Yes, the console has batteries, recently changed, so still pretty strong. 
Rebooting the pi has no effect on the console. it just sits there and 
doesn't undergo any kind of reset/reboot/power cycle.

I had been wondering if my problem is down to the high usage of the 
USB/Ethernet channel. I have three cameras running (each using 
mjpg-streamer to grab/stream frames) plus weewx, also the ethernet data 
rate is high, because the pi's root file system is held on my network NAS 
server. On the other hand, the camera "links" are never broken, it's always 
the WX station, so maybe it's something to do with Python's USB code, or 
this console's handing of USB, or something else ...

I guess I'll go down the syslog-monitoring route suggested by Glenn 
(thanks!) as a reboot once a week-ish isn't a big deal.

Cheers

On Wednesday, October 19, 2016 at 9:57:48 PM UTC+2, mwall wrote:
>
> On Wednesday, October 19, 2016 at 3:38:55 PM UTC-4, Macha wrote:
>>
>> So, 3 Logitech webcams and a WX station. The station is a WEA22 from 
>> WeatherEye (now only £55 from Amazon!). The console is self-powered and 
>> connected directly to USB (no hub).
>>
>
> are there batteries in the weather station console?
>
> when you reboot the pi, does that power cycle the weather station console?
>  
>

-- 
You received this message because you are subscribed to the Google Groups 
"weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to weewx-user+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[weewx-user] Re: Please can I have my bug back?

2016-10-19 Thread mwall 
On Wednesday, October 19, 2016 at 3:38:55 PM UTC-4, Macha wrote:
>
> So, 3 Logitech webcams and a WX station. The station is a WEA22 from 
> WeatherEye (now only £55 from Amazon!). The console is self-powered and 
> connected directly to USB (no hub).
>

are there batteries in the weather station console?

when you reboot the pi, does that power cycle the weather station console?
 

-- 
You received this message because you are subscribed to the Google Groups 
"weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to weewx-user+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[weewx-user] Re: Please can I have my bug back?

2016-10-19 Thread Macha 
Hi,

Thanks for the responses! I'll answer as many questions as I can ..

The RPi (meteopi) running the weather station is part of a small "pi-stack" 
housed in a case. It's powered from a 18V switched mode supply, regulated 
down to 5V by a buck converter. I've checked the voltage on meteopi with a 
scope and it's a tad over 5V, with very little ripple and no "sag". It has 
all USB ports connected. Here's the lsusb output ...

Bus 001 Device 002: ID 0424:9514 Standard Microsystems Corp. 

Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub

Bus 001 Device 003: ID 0424:ec00 Standard Microsystems Corp. 

Bus 001 Device 004: ID 046d:0809 Logitech, Inc. Webcam Pro 9000

Bus 001 Device 005: ID 046d:0809 Logitech, Inc. Webcam Pro 9000

Bus 001 Device 006: ID 1941:8021 Dream Link WH1080 Weather Station / USB 
Missile Launcher

Bus 001 Device 007: ID 046d:0826 Logitech, Inc.

So, 3 Logitech webcams and a WX station. The station is a WEA22 from 
WeatherEye (now only £55 from Amazon!). The console is self-powered and 
connected directly to USB (no hub). One other point worth noting is that 
the root file system is held on my network server, not on the SD card.

The raspbian op system is ...

wheezy - 3.18.7-v7

Python version is 2.7.3 and I seem to have versions 0.1.4 and 1.0.0 of 
libusb installed - presumably the latter is used. pyusb appears to be 0.4.3.

I've attached a copy of my weewx.conf, but the relevant parts are ...

[FineOffsetUSB]

# This section is for the Fine Offset series of weather stations.



# The station model, e.g., WH1080, WS1090, WS2080, WH3081

model = WEA22



# How often to poll the station for data, in seconds

polling_interval = 60



# The driver to use:

driver = weewx.drivers.fousb


[StdArchive]



# If the station hardware supports data logging then the archive 
interval

# will be downloaded from the station. Otherwise, specify it (in 
seconds).

archive_interval = 300



# How long to wait (in seconds) before processing new archive data. Must

# be greater than zero.

archive_delay = 15



# If possible, new archive records are downloaded from the station

# hardware. If the hardware does not support this, then new archive

# records will be generated in software.

# Set the following to "software" to force software record generation.

record_generation = software



# Whether to include LOOP data in hi/low statistics

loop_hilo = True



# The data binding used to save archive records

data_binding = wx_binding

When the "lockup" occurs (perhaps a better description is "lockout"), a RPi 
reboot fixes it, there's no need to do anything to the console, and as 
previously mentioned, it happens about once/twice a week.

I hadn't even begun to think about how I'd monitor syslog, but the rsyslog 
idea is great - I wouldn't have thought of doing it that way ... 

Once again, thanks for the interest, I really appreciate it.


On Wednesday, October 19, 2016 at 10:12:19 AM UTC+2, Macha wrote:
>
> Well, not really I suppose, but ...
>
> Some time ago I commented that my setup (Raspberry Pi + W1080 clone) 
> "crashed" weewx from time to time. I found that restarting weewx didn't 
> help because it couldn't read from the Console. So, I had this cron job 
> which checked if weewx was running. If not, the RPi was rebooted and things 
> worked fine until the next time. This event happened, on average, once a 
> week.
>
> Now, after upgrading to V3.6.0, this kind-of-lockup still occurs, but 
> weewx doesn't crash and simply reports to syslog ...
>
> Caught WeeWxIOError: Cannot determine archive interval
>
>
> ... each 60 seconds.
>
>
> So I still need to "trap" this error in order to reboot the RPi (I have 
> tried many things to "reset" the USB link to the console, but haven't found 
> any solution other than a reboot). The only thing I can think of is to 
> periodically scan syslog and reboot if this message turns up.
>
>
> Is there any other way I can flag weewx errors other than syslog?
>
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to weewx-user+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
# WEEWX CONFIGURATION FILE
#
# Copyright (c) 2009-2015 Tom Keffer 
# See the file LICENSE.txt for your rights.

##

# This section is for general configuration information.

# Set to 1 for extra debug info, otherwise comment it out or set to zero
debug = 0

# Root directory of the weewx data file hierarchy for this station
WEEWX_ROOT = /home/weewx-3.5.0

# How long to wait before timing out a socket (FTP, HTTP) connection
socket_timeout = 20

# Do not modify this. It is used when installing and

[weewx-user] Re: Please can I have my bug back?

2016-10-19 Thread Andrew Milner 
Three further questions - 
1) Is the weather station plugged into a powered usb hub or directly into 
the RPi and 
2) Are you using a beefy power supply for the RPi and/or hub and 
3) if you do have a hub is the hub of the 'correct' type and not 
backfeeding to the RPi


On Wednesday, 19 October 2016 16:14:33 UTC+3, mwall wrote:

> On Wednesday, October 19, 2016 at 4:12:19 AM UTC-4, Macha wrote:
>>
>> Now, after upgrading to V3.6.0, this kind-of-lockup still occurs, but 
>> weewx doesn't crash and simply reports to syslog ...
>>
>> Caught WeeWxIOError: Cannot determine archive interval
>>
>>
>> ... each 60 seconds.
>>
>
> what operating system are you using?
>
> which versions of libusb and pyusb are you using?
>
> how is your weather station console powered?  usb only, or usb and 
> batteries?
>
> please post your weewx.conf file.  specifically, what are your settings 
> for:
>
> [FineOffsetUSB]
> polling_mode = ?
> polling_interval = ?
> [StdArchive]
> archive_interval = ?
> record_generation = ?
>
>
> So I still need to "trap" this error in order to reboot the RPi (I have 
>> tried many things to "reset" the USB link to the console, but haven't found 
>> any solution other than a reboot). The only thing I can think of is to 
>> periodically scan syslog and reboot if this message turns up.
>>
>>
>> Is there any other way I can flag weewx errors other than syslog?
>>
>
> there are many strategies for a nanny.  as glenn noted, you can use syslog 
> itself.  or you can look for the messages from the weewx engine - it 
> retries each 60 seconds because the driver failed, so look for that log 
> message.  or you can look at the latest record in the database to see if it 
> is over an archive interval old.
>
> but first you should be certain of the root cause of the failure.
>
> if you have to power cycle the weather station console to get things to 
> work again, then you are experiencing usb lockup due to firmware problems 
> in weather station console.
>
> you should not have to reboot the pi to restore communication.
>
> if rebooting the pi fixes the comms, and rebooting the pi does not cause 
> the weather station console to power cycle (i.e., you have batteries in the 
> weather station console), then you are experiencing an os/weewx comm 
> failure that can be fixed, probably with modifications to the fousb.py 
> driver.
>
> m
>

-- 
You received this message because you are subscribed to the Google Groups 
"weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to weewx-user+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.