From: =?UTF-8?B?Sm9zw6kgU8OhbmNoZXotR2FsbGVnbw==?= Date: Tue, 10 Jan 2023 09:09:06 -0800 Subject: Re: [sloan-25mlog 6484] 2.5m obslog 20230109 (SJD 59954) --00000000000001574405f1ebf1be Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable > > ---- Software Issues ---- > > (1) Notes to Jose: (a) Two of the designs observed at the end of the nigh= t > belonged to the same field (101131). On the second design, the auto mode > still opted to take a BOSS arc when it shouldn't have taken one since the > telescope was pointed at the same field. (b) Exiting auto mode didn't wor= k. > It started a goto-field sequence for: 12:01:32 Field-Design-Configuratio= n: > 100190-113498-8262. I had tried to toggle off auto mode while the previou= s > exposures were in progress. This happened at the very end of the night. F= or > reference, a time stamp: > > 12:01:45 APO.doravetz_boson 22 hal e error=3D"Goto field failed during au= to > mode." > > Thanks for catching those; I think I fixed both those issues, but please let me know if that's not the case. Right now if you disable the auto mode, the switch in Boson will remain green while the current stage finishes. It's ok to press it multiple times, but it won't get disabled until the macro actually finishes. I'll add something so that's more obvious, like changing it to yellow while it's cancelling. Feel free to try enabling and disabling the auto mode at different times. Hopefully we'll figure out the remaining issues. Also let me know if the RMS target is too conservative and we could be starting the exposure earlier. Right now the acquisition stage stops when it reaches RMS 1 arcsec (and then applies a correction, so the exposure should start with sub-arcsec RMS). Things seem stable-ish enough that I've tagged HAL with those fixes as 0.6.0b1 and it's active at APO. Make sure you use Boson 0.2.1 https://github.com/albireox/boson/releases/tag/v0.2.1 I think that you'll need to install it from the DMG. One of these days I'll really figure out the auto-installer. Regarding the issues with jaeger, I really think there's a networking issue that's making some commands to time out. I don't think that's related to the changes to speed up trajectories; I don't think those sections have changed at all. I've released jaeger 1.3.2 that tries to deal with the issue by increasing the timeouts and doing some retries if the initial command fails, but it's really a workaround. But hopefully it will help a bit. It should also remove the dump of error messages when a command times out. Jos=C3=A9 --00000000000001574405f1ebf1be Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable ---- Software Issues ----(1) = Notes to Jose: (a) Two of the designs observed at the end of the night belo= nged to the same field (101131). On the second design, the auto mode still = opted to take a BOSS arc when it shouldn't have taken one since the tel= escope was pointed at the same field. (b) Exiting auto mode didn't work= . It started a goto-field sequence for: 12:01:32 =C2=A0Field-Design-Configu= ration: 100190-113498-8262. I had tried to toggle off auto mode while the p= revious exposures were in progress. This happened at the very end of the ni= ght. For reference, a time stamp:12:01:45 APO.doravetz_boson 22 hal= e error=3D"Goto field failed during auto mode." Thanks for catching those; I t= hink I fixed both those issues, but please let me know if that's not th= e case.Right now if you disable the auto mode, the switc= h in Boson will remain green while the current stage finishes. It's ok = to press it multiple times, but it won't get disabled until the macro a= ctually finishes. I'll add something so that's more obvious, like c= hanging it to yellow while it's cancelling.Feel free= to try enabling and disabling the auto mode at different times. Hopefully = we'll figure out the remaining issues. Also let me know if the RMS targ= et is too conservative and we could be starting the exposure earlier. Right= now the acquisition stage stops when it reaches RMS 1 arcsec (and then app= lies a correction, so the exposure should start with sub-arcsec RMS).= Things seem stable-ish enough that I've tagged HAL with th= ose fixes as 0.6.0b1 and it's active at APO. Make sure you use Boson 0.= 2.1https://github.com/albireox/boson/releases/tag/v0.2.1I think that you'll need to install it from the DMG. = One of these days I'll really figure out the auto-installer.Regarding the issues with jaeger, I really think there's a net= working issue that's making some commands to time out. I don't thin= k that's related to the changes to speed up trajectories; I don't t= hink those sections have changed at all.I've release= d jaeger 1.3.2 that tries to deal with the issue by increasing the timeouts= and doing some retries if the initial command fails, but it's really a= workaround. But hopefully it will help a bit. It should also remove the du= mp of error messages when a command times out.Jos=C3=A9 --00000000000001574405f1ebf1be--