1.8.183 4k Enova firmware bug?
emdx71
Posts: 50
Loaded the hotfix site firmware for 1.8.183, and while this one boots a ton faster, it looks like its got some serious bugs. In my instance, 'events' don't seem to be firing (for example, I have a scheduled off timer to shut down a display and that no longer fires automatic) also TP feedback events don't seem to be happening. I have a button mute feedback that changes the color of the button from green to red on mute, the button stays on whatever state it was when the system booted. The feedback does get changed on the controller (ie setting the variable from 0 to 1) but that change never makes it to the TP. rolling back to 1.8.171 restores all functionality.
0
Comments
it also appears to be truncating dns names, av.domain.company.com is just domain.company.com
how do you work your schedule timer and the feedback?
Where do you see that dns issue?
DNS showed up because we have a device that is set up to control by DNS name and has worked without issue until this firmware and was throwing errors as unable to connect to dev.domain.company.com when it should be dev.av.domain.company.com, when I went to > Diagnostics >Network Device Addressing the hostname was missing the av. at the beginning (and the controller is set to use DHCP for the address) all other devices on the same network pull the .av at the beginning so its getting set by DHCP properly.
The timer event looks like this:
DATA_EVENT [0:1:0]
{
ONLINE:
{
TIMELINE_CREATE(TL_FDBK,FB_TIME,1,TIMELINE_ABSOLUTE,TIMELINE_REPEAT)
GET_IP_ADDRESS(dvMASTER,IPAddress)
GET_DNS_LIST(dvMASTER,DNSName)
}
}
[TL_FDBK]
{
[vdvTP_Display,251] = [vdvmon,251] // Online
[vdvTP_Display,252] = [vdvmon,252] // Initialized
[vdvTP_Display,255] = [vdvmon,255] // Power State
IF (DAY == 'MON' OR DAY == 'TUE' OR DAY == 'WED' OR DAY == 'THU' OR DAY == 'FRI' AND TIME > '06:00:00' AND TIME <'06:01:00')
{
PULSE [vdvmon,27]
}
ELSE IF (DAY == 'MON' OR DAY == 'TUE' OR DAY == 'WED' OR DAY == 'THU' OR DAY == 'FRI' AND TIME > '15:30:00' AND TIME <'15:31:00')
{
PULSE [vdvmon,28]
}
}
}
That feedback for the monitor module state doesn't work either, nor does power state feedback. Its almost as feedback isn't firing in this build. if I roll back all works fine.
I should also clarify the version of firmware I rolled back to was 1.8.123 which made everything work again
Regarding the DNS name, I set it from Studio -> Diagnostic -> Network addresses.
Working as expected.
The result of the attached code (time difference comes because the NX is a minute ahead from my computer
)
The time check works, and the pulse was done, so also the timeline is running. Also emulating vdvmon channel 251 on/off will update [vdvTP_Display,251]
Had no Varia on hand, but can give a try after weekend
With the code in that form, because of the 1 minute timeframe the condition is true, the pulse would be done 60 times within that minute. But basically it works.
Rename the attached TXT file back to axs, to see my test program.
I'll try it out, for the DNS, you can set the name and it will be fine, the issue comes when the name is coming from dhcp. it shortens what is provided by dhcp. the name provided by dhcp was av.domain.company.com (Made up name for this example) and gets shortened to domain.company.com. tested 2 differnet units, a an nx1200 and a 2265 and both yielded the same results.
set up my network with a DHCP (using a Netgear AV line switch as dhcp server) that provides the domain "dev.av.domain.company.com", works fine.
And with another Domain Suffix "netlinx.av.support.audiopro.de" provided by the DNS server
Show DNS List
NX3200 v1.8.183
timelines aren't running for me with this build. downloaded a new copy as well, and loaded to 3 different controllers. all my timeline variables that are flagged when running are stuck at 0.
also events that are tied to timelines via button presses don't work either. I have a timeline that cycles an access key, that stays either at 0 or blank when the button is pressed.
update, I did a clean disk -f duet on one of the controllers that wasn't working, cleared all the persistent variable and now it seems to be firing feedback.
Update 2:
Physically power cycled the controller and feedback is again not running, rebooting twice after the power cycle and feedback runs again
not reproducable by me.... find attached my test program, rename it from txt to axs.
Buttons 1 starts a relative timeline
Buttons 2 start an absolute timeline
Button 3 kills both timeline
sequence numbers will be given out to diagnostics while running.
There are 2 status variables for these timelines, manually set at timeline_create (TRUE) and timeline_kill (FALSE)
A 3rd timeline will watch the status of the above timelines in the way of a typical feedback timeline.
In this timeline_event, the status of the above 2 timelines is checked and updated by timeline_active() (timeline_active() reports if a timeline is currently existing in the system (result is the ID of the timeline checked) or not (result is 0) )
I am also having an issue with timelines not starting on newer 1.8 firmware. (I have 1.6.205 on most NX-x200 controllers right now.) Doing a clean disk made it work, but, I'm worried it will come back.
emdx71, did you find a long term solution?
Solving by Clean Disk sounds like a Persistent Variable is blocking something. Because Persistent Vars keep their values if they are defined in a new program download, you can't solve this without a Clean Disk, as long as such variable is not reseted by Define_Start...
The latest hotfix 1.8.200 seems to have several improvements, worth a try...
Not one that works 100%. I've been keeping 1.6.205 in production. I rolled out a 1.8.something (193 or 196 I can't remember) to a production room and after about a week the system totally froze. I reached out to support and their suggestion was to attach the TIMELINE_CREATE on a the online event of a virtual device. I built a variable called nFB_Running and rather than a virtual device, I attached it to the online event for the touch panel. I created an IF statement for nFB_running and in the data if that variable isn't 1, there is a timeline_create there for feedback & the timeline is supposed to start. Notice I said supposed to. That doesn't even always work. Its been more consistent, but only about 80 to 90%. When the timeline won't start, nothing will make it start. I've even assigned it to a button to start it. When it won't start, your only recourse is to restart. I made an event for that as well. that if the timeline isn't running and after a 4 minute wait still isn't, then reboot 0:1:0. But then the the box doesn't always come back after a reboot and a power cycle is needed. I've had the same issue with all the 1.8s even the .200. I was hopeful this morning when it rebooted quicker, but then timelines were't running. a reboot got them going though.
I did some digging on the SDcards and it looks like one of the major differences between 1.6 and 1.8 is 1.6 uses a 2.6x linux kernel and 1.8 looks like it uses a 4.x kernel. I'm guessing with the kernel changes newer packages for java are needed (hence the security path changes) and something isn't getting kicked off in the init process. but that's purely speculation on my part.
The one thing I've noticed a lot more is "timed out waiting for duet" shows up always at boot in diag, but it references the ICSLan which we don't use so I suppose thats expected.