I imported this release into virtual-box. Set it up same way as 1.6.0 but it just sits on Ubuntu 14.04 . . . . I've rebooted it a few times and still have same issue... I can leave it there for over an hour and it will still be loading...
I think the Spectre/Meltdown patch must have slowed this thing to a crawl.
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
3790 shhh 20 0 4708192 942736 832324 S 239.9 3.5 34:05.23 VirtualBox
239% CPU load....
After awhile I gave up. Reloaded 1.6.0 back onto it and pushed backup and back to normal.
Hi Rynet -- is the issue you are unable to load the image into VirtualBox and start it? I just downloaded and installed it on my laptop using VirtualBox, and I was able to launch and log in fairly easily with the default settings. Else, is the issue you installed it, but are unable to import your previous environment and get it working? If that is the case, can you please let us know what your VM configuration is, and what device count you are managing? It would also be helpful to know the breakdown of device types.
After I imported the appliance .ova it would just sit on the screen saying ubuntu . . . and eat up cpu cycles.
Didnt even get as far as to set its IP or log in to GUI to import backup.
Hi Rynet -- I am still not seeing the issue on startup or reboot. I am using Virtual Box 5.2.8 r121009 (so the latest). [I am editing the second half of this message, because there was an error in my initial reply.] The issue may be a side effect of an increase in base memory from 2048 MB (which is what it has been since inception) to 4096 MB. This was bumped up, because additional features and database updates have caused RAM use to increase in the base system, which was leading to issues in some smaller customer environments. I will update the Release Notes to be clear on this later today. Could this be a possible issue with your environment?
Waiting for the "Clearing Major Alarms" feature
http://community.cambiumnetworks.com/t5/Your-Ideas/cnMaestro-Clear-Major-Alarms/idi-p/82738
VLAN ID in RestfulApi has been implemented?
First - Don't forget to make this a "Pinned" thread.
Second - cnMaestro on-premises
SM "Session Time" not being read from the device.
Same SM
Hi Luis,
We verified session time for both ePMP/PMP SMs in our lab setup and it seems to be working fine. Could you please email tech support file from problematic SM to me - ajay.singh@cambiumnetworks.com for further investigation.
Thanks,
Ajay
edit_______________
I obviously didn't read your whole reply. Just the part where you "verified" the problem. Also I thought the two kudos meant they also had the problem.
Edit #2 ____________
My vanity got me!!!
I didn't even read your reply - I was wondering how the heck I misread it so bad. I saw the message indicator and it took me to the "Kudos" page and I replied from there.
_______________________________
_______________________________
This is what I ment by percolate.
I was hoping these obvious things would be caught by those using the cloud version. I know they update it on the fly while the on-premises updates are a bit slower. I know this is rev 10 where the cloud is 9 so maybe it was something introduced in that update.
I'm still seeing a problem with UTC -vs- Local time in statistics and graphing data. Also data missing altogether, Uplink RSSI and Uplink RSSI Imbalance comes to mind.
I'll report them in detail later.
Also, thanks for adding the description field to the AP and SM widgets and all the team's work in making cnMaestro a better and more useful tool.
@Cambium_Ajay wrote:Hi Luis,
We verified session time for both ePMP/PMP SMs in our lab setup and it seems to be working fine. Could you please email tech support file from problematic SM to me - ajay.singh@cambiumnetworks.com for further investigation.
Thanks,
Ajay
It's not just one SM, its all. Did you verify by doing an export, update and import of cnMaestro or on a clean run? Though I don't see why that would make a difference if it's a live read. Having said that, I didn't think to add this fact as the times don't correlate with the update time.
Okay. Indeed, this is strage that all SMs are reporting incorrect session time in cnMaestro. We validated on a running system and not really with data export and import. Looks very unlikely that data export/imprt will cause this issue as we read session time from PMP SMs itself. Anyway we will validate this scenario as well and let you know.
-Ajay
@Cambium_Ajay wrote:Okay. Indeed, this is strage that all SMs are reporting incorrect session time in cnMaestro. We validated on a running system and not really with data export and import. Looks very unlikely that data export/imprt will cause this issue as we read session time from PMP SMs itself. Anyway we will validate this scenario as well and let you know.
-Ajay
Supposed to, anyway. It may not be the read that's wrong, it's probably just the display. I remember reading a thread about incorrect display of a time field on a report export, but can't remember which. It wasn't even formatted correctly if I remember. It was in seconds UTC, I believe. I'll have to take a look.
I know the session time display was wrong, then corrected and now wrong again. I just don't remember which rev corrected the display. I'll look for this also.
Edit____________
https://community.cambiumnetworks.com/t5/cnMaestro/Problem-with-export/m-p/84458#M3084
After running for two days - no correct cnMaestro session time display for any of the SMs.
Rebooted an SM to see what would happen. Here's the result.
Now the read is correct - Where the disconnect is before the reboot, I don't know. I do know I'm not going to reboot the entire network just to get cnMaestro to display the data correctly. I'll take this opportunity to ask about the arrival of PMP 15.2.
Before you tell me you couldn't duplicate the error:
Run an instance of cnMaestro on-premises 1.6.0 r22
Onboard SM (I'm using 450 vanilla), prefferably one that's been up for a while and don't turn it off til the end
Let it run long enough for cnMaestro to collect session time data
Verify session time between the two
Export cnMaestro data
Install cnMaestro 1.6.1 r10
Import data
Let SM run long enough to update cnMaestro with new session time
Verify session time - this is when I see the discrepancy
If you were able to duplicate, which you should be able to
reboot SM and wait for session time to update
Verify session time - times should now be the same
Well?
I upgraded Virtual Box from 5.1 to 5.2.8 and it fixed the issue. The issue is that with the new Linux Kernel in Ubuntu and probably other releases is that using VT-x/AMD-V without a Nested Paging file causes virtual machines to crash.
This is an issue with older processors. As they cannot do a Nested Paging File correctly. But in 5.2.8 this has been fixed.
Hi Rynet -- glad you were able to figure it out and get it working. We will update our User Guide for Virtual Box versioning.
Thankfully, this all was not needed and we could find the root cause. So in short, there was an issue in the software, wherein we weren’t parsing the day part of the session time string received from PMP SM device. So any PMP SM device which is running for more than 24 hours will have incorrect session time in cnMaestro. We have prioritized the fix and same will be available in upcoming release – 1.6.2, tentatively scheduled for release by 16-March. Hope this explains the current issue.
-Ajay
@Cambium_Ajay wrote:... We have prioritized the fix and same will be available in upcoming release – 1.6.2, tentatively scheduled for release by 16-March. Hope this explains the current issue.
-Ajay
This is great.
Thanks!
On that note...
3/6 - I set up two reports to run on a daily basis - "Performance" and "Events". I also ran them as "Now" reports to get that day's data.
The "Now" reports ran without issues.
3/7 - Both scheduled reports ran and the data was as I had expected. Super! I wanted to modify the times the reports would run, but once a scheduled report runs, it can't be modified as the "Edit" icon is disabled. No problem I guess, I just deleted the two jobs and set them up anew. Checked to make sure they were set up in "Jobs" and there they were. While there, I noticed the job "ID" didn't reset. The "Software Update" job "ID" resets, so maybe a standardization problem here.
3/8 - Alas, the reports either didn't run or they retrieved the same data as the day before. There's no way to tell as there's no run counter or any other indication like a date on the report or the "Job" dashboard. When I set the jobs up, I noticed one of the criteria is "End By" not "End On". If the program gets to pick when to quit, then I guess it's working just fine. ;)
I set both reports to export daily, one starting 3/7 at 09:00 and ending by 12/31 at 10:00 the other one starting 3/7 at 09:05 and ending by 12/31 at 11:15.
I haven't tried running after x occurrances.
If I set a daily job and come back after several days - a weekend, I'm expecting there will be three different reports waiting to be downloaded - Saturday, Sunday and Monday. Am I wrong in this expectation? How will subsequent reports be handled? Do they just append each other? As I type this, I'm remembering that I haven't read the new user guide. My apologies if it's referenced there.
Also, "System/Statistics" doesn't export the "Status" time.
At a minimum, the "Status" time" should be exported in the "Status" cell. It would be better, IMHO, if both the MAC address and "Status" time were in their own cells.
None of the reports export the "Session Time", don't remember about the "Status" time.
Thanks again.
Adding a post instead of editing.
The program must have known that I was reporting it's misbehavior! The jobs ran while I was typing the post. If you notice on the snapshot of the "Jobs" page, I ran a "Now" job (#7) at 09:45 when I found the reports had not run. Oh well!
I was able to export them both by direct "Download" and by downloding from the "History" page, thus answering the question about how subsequent reports are handled. I guess the last report run will be available by direct "Download" from the "Jobs" page.
Thanks again
"Statistics" page not displaying "Uplink RSSIIImbalance" nor "Uplink RSSI". Maybe it's already being imported from the AP along with "Modulation (UL)" and just not being displayed.
It would be great if you added them after "Modulation (UL), in the listed order, and label both the current RSSI fields with (DL).
"Performance" report not exporting "Uplink RSSI (dBm)" nor "Uplink RSSIImbalance"
This is how I use the data.
Thanks
