Re: [Nut-upsuser] Powerwalker VI 2200VA LCD

[hyo...@gmail.com]

Lars, the log ends around 41998 seconds (~ 11 1/2 hours) and, till
that point, from what I can see, nothing strange happens: both the
interrupt pipe and polling work as expected.
Now, you mentioned that something changes when you alter the polling
frequency (I think you refer to 'pollfreq' and not to 'pollinterval',
but correct me if I'm wrong), but without seeing it happen, we can't
say whether the problem is caused by an interrupt, a quick update or a
full update.

Another couple of things you could try to narrow down the candidates:
- disable the interrupt pipe ('pollonly'), or
- (if you can upgrade to at least 2.7.4) disable polling ('interruptonly').

___
Nut-upsuser mailing list
Nut-upsuser@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser

Re: [Nut-upsuser] Powerwalker VI 2200VA LCD

[Lars de Bruin]

I dont think the logs ends to early.
The UPS had been offline for 1 day at least before i stopped the process.
Could the issue be that NUT driver does not see the USB device disconnect?






 Original Message 
Subject: Re: [Nut-upsuser] Powerwalker VI 2200VA LCD
Local Time: 11 November 2016 1:12 AM
UTC Time: 11 November 2016 00:12
From: hyo...@gmail.com
To: Lars de Bruin <l...@larsdebruin.net>
nut-upsuser <nut-upsuser@lists.alioth.debian.org>

2016-11-09 14:57 GMT+01:00 Lars de Bruin <l...@larsdebruin.net>:
> Hi,
>
> - The device never re-connects to the system, i have to physically remove
> the USB plug and insert it again for the device to come back.
> - This does not happen when running the driver from the supplier. (java
> process)

Ok, if they do, we should be able to do it too.

> - When i poll every 60s it works for about 3 days, anything lower or higher
> causes it to disconnect faster
> - The big log i sent is not the output of the driver? if not how can i
> enable that ?

Sorry if I wasn't clear, it is the right log, but it ends too early:
we could get more answers if the log covers the moments before the
disconnection.

>
>
> Thank you for your time..___
Nut-upsuser mailing list
Nut-upsuser@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser

Re: [Nut-upsuser] Powerwalker VI 2200VA LCD

[Stuart Gathman]

On 11/10/2016 07:12 PM, hyo...@gmail.com wrote:
> 2016-11-09 14:57 GMT+01:00 Lars de Bruin :
>> - The device never re-connects to the system, i have to physically remove
>> the USB plug and insert it again for the device to come back.
This is my log with a Tripplite SMART1500LCDT on CentOS-6 with
nut-2.6.5, also disconnecting with the same timing:

Nov  6 09:54:18 fairfax kernel: hub 1-1.1.1:1.0: port 4 disabled by hub (EMI?), 
re-enabling...
Nov  6 09:54:18 fairfax kernel: usb 1-1.1.1.4: USB disconnect, device number 14
Nov  6 09:54:21 fairfax named[28979]: transfer of 'neo/IN' from 
2002:cfc0:470d::1#53: failed to connect: timed out
Nov  6 09:54:21 fairfax named[28979]: transfer of 'neo/IN' from 
2002:cfc0:470d::1#53: Transfer completed: 0 messages, 0 records, 0 bytes, 
63.000 secs (0 bytes/sec)
Nov  6 09:54:22 fairfax kernel: hub 1-1.1.1:1.0: Cannot enable port 4.  Maybe 
the USB cable is bad?
Nov  6 09:54:26 fairfax kernel: hub 1-1.1.1:1.0: Cannot enable port 4.  Maybe 
the USB cable is bad?
... 
Nov  6 09:54:34 fairfax kernel: hub 1-1.1.1:1.0: unable to enumerate USB device 
on port 4

No, the cable is not bad. 

While not really solving the problem, I was able to work around it by
removing power from the USB port under software control.  This keeps the
UPS online while unattended.

https://gathman.org/2016/07/30/Standard_Schmandard/
https://github.com/sdgathman/trippfix

If you have a hub with power control (list here:
https://github.com/sdgathman/trippfix#hubs-known-to-work), you can give
it a try.
The package uses upssched to append to a file, and incrond to then run a
script to cycle the port power as root.  I switched to hub-ctrl from
hubpower because the former has a GPL license.

Here is a log of the disconnects:

2016-08-27 23:23:56 NOCOMM
2016-08-30 14:38:21 NOCOMM
2016-08-30 15:25:34 NOCOMM
2016-08-30 15:41:40 NOCOMM
2016-09-02 08:41:37 NOCOMM
2016-09-04 02:08:26 NOCOMM
2016-09-09 23:57:25 NOCOMM
2016-09-10 17:02:54 NOCOMM
2016-09-10 22:30:40 NOCOMM
2016-09-15 04:36:36 NOCOMM
2016-09-15 19:43:25 NOCOMM
2016-09-17 20:21:20 NOCOMM
2016-09-19 13:54:36 NOCOMM
2016-09-19 14:00:59 NOCOMM
2016-09-19 15:03:51 NOCOMM
2016-09-22 19:10:20 NOCOMM
2016-09-22 22:41:21 NOCOMM
2016-09-23 20:06:56 NOCOMM
2016-09-30 00:12:17 NOCOMM
2016-09-30 16:43:56 NOCOMM
2016-10-01 03:16:23 NOCOMM
2016-10-02 22:20:32 NOCOMM
2016-10-06 23:11:42 NOCOMM
2016-10-10 02:03:32 NOCOMM
2016-10-12 17:24:27 NOCOMM
2016-10-14 06:32:10 NOCOMM
2016-10-16 20:23:02 NOCOMM
2016-10-16 21:47:42 NOCOMM
2016-10-21 01:04:48 NOCOMM
2016-10-23 05:07:29 NOCOMM
2016-10-23 13:32:40 NOCOMM
2016-10-26 06:43:35 NOCOMM
2016-10-26 13:03:26 NOCOMM
2016-10-26 17:43:27 NOCOMM
2016-10-28 21:19:33 NOCOMM
2016-10-29 06:10:25 NOCOMM
2016-10-30 23:08:29 NOCOMM
2016-11-02 11:58:55 NOCOMM
2016-11-04 11:31:14 NOCOMM
2016-11-06 09:54:44 NOCOMM
2016-11-09 01:45:53 NOCOMM
2016-11-09 08:47:48 NOCOMM
2016-11-09 17:02:52 NOCOMM



___
Nut-upsuser mailing list
Nut-upsuser@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser

Re: [Nut-upsuser] Powerwalker VI 2200VA LCD

[hyo...@gmail.com]

2016-11-09 14:57 GMT+01:00 Lars de Bruin :
> Hi,
>
> - The device never re-connects to the system, i have to physically remove
> the USB plug and insert it again for the device to come back.
> - This does not happen when running the driver from the supplier. (java
> process)

Ok, if they do, we should be able to do it too.

> - When i poll every 60s it works for about 3 days, anything lower or higher
> causes it to disconnect faster
> - The big log i sent is not the output of the driver? if not how can i
> enable that ?

Sorry if I wasn't clear, it is the right log, but it ends too early:
we could get more answers if the log covers the moments before the
disconnection.

>
>
> Thank you for your time..

___
Nut-upsuser mailing list
Nut-upsuser@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser

Re: [Nut-upsuser] Powerwalker VI 2200VA LCD

[Lars de Bruin]

Hi,

- The device never re-connects to the system, i have to physically remove the 
USB plug and insert it again for the device to come back.
- This does not happen when running the driver from the supplier. (java process)
- When i poll every 60s it works for about 3 days, anything lower or higher 
causes it to disconnect faster
- The big log i sent is not the output of the driver? if not how can i enable 
that ?



Thank you for your time..


 Original Message 
Subject: Re: [Nut-upsuser] Powerwalker VI 2200VA LCD
Local Time: November 9, 2016 12:54 AM
UTC Time: November 8, 2016 11:54 PM
From: hyo...@gmail.com
To: Lars de Bruin <l...@larsdebruin.net>
nut-upsuser <nut-upsuser@lists.alioth.debian.org>

The huge log doesn't show anything troublesome or problematic.

The part of the driver output you posted on your third-to-last email
only shows that the device disconnects and then it doesn't reconnect,
or, at least, the driver doesn't see it:
- At that point, does the device reconnet to the system? (Does it show
up on dmesg? What does `lsusb -d 06da: -v` return?)
- It could be helpful to know what happened just before the
disconnection to see if there is any correlation with driver activity:
any chance you have the log of the driver running in debug mode till
that point?

A fair guess could be that the device is already busy doing something
and considers communication of a lower priority and hence stops it.
Were this the case (but, then, I would expect the device to reconnect
and the driver to restart communication), a 'full update' would likely
be the culprit, and hence increasing 'pollfreq', while not curing the
problem, could at least decrease the chance of experiencing it.___
Nut-upsuser mailing list
Nut-upsuser@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser

Re: [Nut-upsuser] Powerwalker VI 2200VA LCD

[hyo...@gmail.com]

The huge log doesn't show anything troublesome or problematic.

The part of the driver output you posted on your third-to-last email
only shows that the device disconnects and then it doesn't reconnect,
or, at least, the driver doesn't see it:
- At that point, does the device reconnet to the system? (Does it show
up on dmesg? What does `lsusb -d 06da: -v` return?)
- It could be helpful to know what happened just before the
disconnection to see if there is any correlation with driver activity:
any chance you have the log of the driver running in debug mode till
that point?

A fair guess could be that the device is already busy doing something
and considers communication of a lower priority and hence stops it.
Were this the case (but, then, I would expect the device to reconnect
and the driver to restart communication), a 'full update' would likely
be the culprit, and hence increasing 'pollfreq', while not curing the
problem, could at least decrease the chance of experiencing it.

___
Nut-upsuser mailing list
Nut-upsuser@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser

Re: [Nut-upsuser] Powerwalker VI 2200VA LCD

[Lars de Bruin]

Was anybody able to read the logs and come up with an error?






 Original Message 
Subject: Re: [Nut-upsuser] Powerwalker VI 2200VA LCD
Local Time: October 16, 2016 11:47 PM
UTC Time: October 16, 2016 9:47 PM
From: hyo...@gmail.com
To: Lars de Bruin <l...@larsdebruin.net>
nut-upsuser <nut-upsuser@lists.alioth.debian.org>

2016-10-10 15:28 GMT+02:00 Lars de Bruin <l...@larsdebruin.net>:
> The UPS seems to disconnect totally, i need to physically disconnect the
> cable to get the UPS connected again.
> A reboot is not enough.

Ouch, this is more severe than what I thought.
As a first step, you could try and increase the timeouts for pollfreq
and pollinterval.
Again, a log (compressed) of the driver running in debug mode when
those problems occur would be helpful.___
Nut-upsuser mailing list
Nut-upsuser@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser

Re: [Nut-upsuser] Powerwalker VI 2200VA LCD

[Lars de Bruin]

I have run nut driver in debug level 6.
Here is the log: http://www.filedropper.com/ups_1

Please have a look, i cannot see a pattern since i dont know what to look for.
DMESG:

[269391.234426] hid-generic 0003:06DA:.0003: hiddev0,hidraw0: USB HID v1.00 
Device [PPC Offline UPS] on usb-2098.usb-1.2/input0
[285040.957004] usb 1-1.2: usbfs: USBDEVFS_CONTROL failed cmd usbhid-ups rqt 
128 rq 6 len 255 ret -75





 Original Message 
Subject: Re: [Nut-upsuser] Powerwalker VI 2200VA LCD
Local Time: October 16, 2016 11:47 PM
UTC Time: October 16, 2016 9:47 PM
From: hyo...@gmail.com
To: Lars de Bruin <l...@larsdebruin.net>
nut-upsuser <nut-upsuser@lists.alioth.debian.org>

2016-10-10 15:28 GMT+02:00 Lars de Bruin <l...@larsdebruin.net>:
> The UPS seems to disconnect totally, i need to physically disconnect the
> cable to get the UPS connected again.
> A reboot is not enough.

Ouch, this is more severe than what I thought.
As a first step, you could try and increase the timeouts for pollfreq
and pollinterval.
Again, a log (compressed) of the driver running in debug mode when
those problems occur would be helpful.___
Nut-upsuser mailing list
Nut-upsuser@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser

Re: [Nut-upsuser] Powerwalker VI 2200VA LCD

[Lars de Bruin]

 not claim 
interface 0 before use


[160058.243044] usb 1-1.2: usbfs: process 334 (usbhid-ups) did not claim 
interface 0 before use


223920.728609] usb 1-1.2: USB disconnect, device number 5






 Original Message 
Subject: Re: [Nut-upsuser] Powerwalker VI 2200VA LCD
Local Time: October 16, 2016 11:47 PM
UTC Time: October 16, 2016 9:47 PM
From: hyo...@gmail.com
To: Lars de Bruin <l...@larsdebruin.net>
nut-upsuser <nut-upsuser@lists.alioth.debian.org>

2016-10-10 15:28 GMT+02:00 Lars de Bruin <l...@larsdebruin.net>:
> The UPS seems to disconnect totally, i need to physically disconnect the
> cable to get the UPS connected again.
> A reboot is not enough.

Ouch, this is more severe than what I thought.
As a first step, you could try and increase the timeouts for pollfreq
and pollinterval.
Again, a log (compressed) of the driver running in debug mode when
those problems occur would be helpful.___
Nut-upsuser mailing list
Nut-upsuser@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser

Re: [Nut-upsuser] Powerwalker VI 2200VA LCD

[hyo...@gmail.com]

2016-10-10 15:28 GMT+02:00 Lars de Bruin :
> The UPS seems to disconnect totally, i need to physically disconnect the
> cable to get the UPS connected again.
> A reboot is not enough.

Ouch, this is more severe than what I thought.
As a first step, you could try and increase the timeouts for pollfreq
and pollinterval.
Again, a log (compressed) of the driver running in debug mode when
those problems occur would be helpful.

___
Nut-upsuser mailing list
Nut-upsuser@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser

Re: [Nut-upsuser] Powerwalker VI 2200VA LCD

[Lars de Bruin]

The UPS seems to disconnect totally, i need to physically disconnect the cable 
to get the UPS connected again.
A reboot is not enough.
Also tried 2 different cables, the one i use now is shielded + 2 ferrite chokes 
to limit interference.


But i want to use NUT since i dont have enough resources to run X..




 Original Message 
Subject: Re: [Nut-upsuser] Powerwalker VI 2200VA LCD
Local Time: October 10, 2016 12:23 AM
UTC Time: October 9, 2016 10:23 PM
From: hyo...@gmail.com
To: Lars de Bruin <l...@larsdebruin.net>
nut-upsuser <nut-upsuser@lists.alioth.debian.org>

2016-10-03 14:56 GMT+02:00 Lars de Bruin <l...@larsdebruin.net>:
> Hi,
>
> The UPS is now connected via HID-USB, but seems to randomly disconnect.
> Sometimes after a few hours, other times days...
> Any suggestions?

A log would help, but, in general, if the device disconnects and,
after a short time, the connection is restored, it could be a cable
problem, or the UPS not keeping the pace needed to update all the
necessary data.___
Nut-upsuser mailing list
Nut-upsuser@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser

Re: [Nut-upsuser] Powerwalker VI 2200VA LCD

[hyo...@gmail.com]

2016-10-03 14:56 GMT+02:00 Lars de Bruin :
> Hi,
>
> The UPS is now connected via HID-USB, but seems to randomly disconnect.
> Sometimes after a few hours, other times days...
> Any suggestions?

A log would help, but, in general, if the device disconnects and,
after a short time, the connection is restored, it could be a cable
problem, or the UPS not keeping the pace needed to update all the
necessary data.

___
Nut-upsuser mailing list
Nut-upsuser@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser

Re: [Nut-upsuser] Powerwalker VI 2200VA LCD

[Lars de Bruin]

Hi,

The UPS is now connected via HID-USB, but seems to randomly disconnect.
Sometimes after a few hours, other times days...
Any suggestions?




 Original Message 
Subject: Re: [Nut-upsuser] Powerwalker VI 2200VA LCD
Local Time: October 2, 2016 11:50 PM
UTC Time: October 2, 2016 9:50 PM
From: hyo...@gmail.com
To: Lars de Bruin <l...@larsdebruin.net>
nut-upsuser <nut-upsuser@lists.alioth.debian.org>

> [powerwalker]
> driver = blazer_usb
> vendorid = 06da
> port = auto
> desc = "Powerwalker VI 2200VA LCD"
> productid = 
> bus = "001"

The VID/PID combo 06da/ is used by some devices supported by the
'usbhid-ups' driver: did you have a chance to try it?___
Nut-upsuser mailing list
Nut-upsuser@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser

Re: [Nut-upsuser] Powerwalker VI 2200VA LCD

[hyo...@gmail.com]

> [powerwalker]
>driver = blazer_usb
>vendorid = 06da
>port = auto
>desc = "Powerwalker VI 2200VA LCD"
>productid = 
>bus = "001"

The VID/PID combo 06da/ is used by some devices supported by the
'usbhid-ups' driver: did you have a chance to try it?

___
Nut-upsuser mailing list
Nut-upsuser@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser