Last nights session was rough as something changed. The RPI3 kept running out of memory so either Kstars or PHD2 would be killed. I am looking to see if creating a swapfile/zram, as well as turning off all debugging for tonights session to see if it helps reduce footprint...

Additionally:

- Ekos is not connecting to PHD2 - the port, 4400 I think, is not being saved and is always reverting to 0. I can run PHD2 separately, it is just that Ekos cannot connect to control it.

- Filter offsets do not seem to be saved. They are set in the CCD module and are always zero... Not sure why I have wrong. I will try revising the NightCrawler INDI driver settings tonight.

- On the CCD tab, are the autofocus, refocus, meridan flip, guide deviations, and Fits viewer settings saved as part of the CCD SEQUENCE

- Is the park position saved in the mount settings? It seems to get reset to 0's allot.

- Is there a KStar notification for morning twighlight?

Thank you

+ What process was consuming RAM? Use top or htop to find out.
+ For PHD2, there was a typo in the userdb.sqlite schema earlier that was fixed. So you have two options: edit ~/.local/share/kstars/userdb.sqlite and then edit the profile table, check if there is a column call guiderpost. If it exists, rename it to guiderport and save.
+ The offsets are also saved in the userdb.sqlite (filter table). So check those as well. Maybe it is best if you attach your userdb.sqlite
+ Park position is saved in ~/.indi/ParkData.xml ... the zeros reported is a bug that I have to track down since it's dangerous.
+ No notification unless you're using scheduler and then it will shutdown. But adding another warning would be nice, thanks for the suggestion!

I found out the issue with filters table, it will be fixed in the next release 2.8.3

- I will use htop next time. The session was better with debugging turned off on the indi-drivers. A quick check suggested it may be a specific a driver that had the most memory, and I will confirm with htop next session before suggesting which one.

- found and fixed the column to guiderport - Thank you, all is good now.

- Do you still want my userdb.sqlite?

- The park may be a user error. Do I need to click the Write Data button? After setting the Home Pos with CGEM HC in last session, in the INDI driver I clicked the CURRENT button and then Write Data button. The park position seemed to work last session. I check the value just now with Simulation on and it seems to be the same values.


- I am trying to get the scheduler and PHD2 to play nice come morning twighlight. Ekos shut down, but I have not gotten the two to behave in sync yet, and PHD2 seems to continue looping. The twighlight notification would allow the running of this script come twighlight to shut down PHD2 if it is still running:

#!/bin/sh
nc localhost 4400 <<EOF
{"method": "stop_capture"}
EOF
#****

It is cloudy now for a awhile, so it can be a some days before I can update again. If I can test in house I will.

Thank you, Cheers!

No need for userdb.sqlite.

If you use current, and then issued park command then no need to click "write data" as it is called automatically. Regarding Scheduler, it should abort guiding including PHD2 when it stops.

On the CCD tab it is clear that the sequence for filters and exposures are saved in the ESQ file.

When and where are the other settings saved. It is not clear if it is saved per line in the sequence, per the whole ESQ file, not saved as part of the sequence file at all, or cannot be changed once the sequence, or schedule is running.

As an example, consider a scheduled sessions is running. There is a sequence file with one 120s exposure per LRGB filters - four entries. The scheduler is set to repeat until twighlight, and is running and the first LUM exposure.

If the meridian setting is toggled on or off on the CCD tab does it affect anything? Same thing for the Enable Fits viewer and Auto Stretch, to changing the minutes for the refocus, what happens if these are changed during a running schedule?

I am trying to understand if I change the enable fits viewer, or changing the time for the meridian flip, or focuser, what do I need to do for it to take effect.

Thank you

Ok so it's fairly simple:
+ Job changes are not reflected unless you edit the job, edit the values, and commit the changes it to it (by pressing the checkmark). So randomly changing stuff there have no effect on the job.
+ Universal settings like limit settings (meridian flip..etc) can be changed in run time.

Working on a imaging session now. For the Fits Viewer enable universal, does it have to be unchecked before the image is started? I asked as I am running into memory issues.

It is the indi_asi_ccd that seems to be consuming the most memory. Some check via top showed it as small as 4.4% at start and routinely as 17% on RPI3.

I have been working for a few hours to get the framing and focus and imaging going. It would be an amazing session tonight except (final clear even though almost a full moon) if it were not for either indi, phd2, or kstars being killed.

I tried to even use Kstars on mac and still it seems something is consuming memory as indi or PHD 2 would then be killed.

You should probably enable Limited Resources Mode under KStars Settings ---> FITS to reduce memory usage in KStars. For ASI driver, I haven't seen this before I'd have to check.

1. Another PHD2 item tonight is that I cannot stop PHD2 from sending the guide frame onto Kstars. The setup tab summary continually shows the guide frame. The light frame will appear until it is replaced. Unchecking the IMAGE data in INDI for the ASI174 camera was tried too to no avail.

2. Kstars could not run on the RPI3 tonight. The night may be saved though as the third time (finally have one frame of NGC6888) seems to be the charm to get the Mac Kstars connecting okay. It is the first time I am using the Mac for the session and the RPI3 as the controller only. I am working on the Ekos VM as a backup, solution to access the RPI3 as the indiserver. It was just started though and not possible to do tonight.... So ya, Mac seems to be work. Not ya, my RPI3 could not do it all tonight.

The majors symptoms tonight is memory being maxed with either Kstars, INDI, or PHD2 being killed on RPI3.

There is another symptom and that is the indiwebserver would not work, no connections tonight. Not sure why as the servermanager.py (forget the spelling) is running, but nothing is connected.

Also, it seems Ekos is connected with with PHD2 tonight. If all goes right PHD2 should be stopped with the session tonight. This will be good for me as it is typical that the sessions stop with twighlight on time, but PHD2 keeps going.

I will get the logs I have and attach them tomorrow (I disabled some logging hoping it would help with memory)

I do have the Limited Resources Mode enabled.

Another item, is that the ASI_EFW was not saving its filter names. Whether the filters names were saved on the RPI3 or Mac version of Ekos, the names reverted to Filter #1...7.

This is odd as Ekos was just used days ago.

I am certain there is something weird I had done.

Perhaps it has something to do with using the local indiserver tonight. I have always used the Web indi server, and where it would not work I tried local. It is worth noting that the ASI-EFW never used to start and I had to use indiwebserver.

I did to an update yesterday and today - sudo apt-get update and sudo apt-get dist-upgrade