Forum Discussion
Spatial Anchor Cloud Save Fails
Hello, I'm trying to use "Shared Spatial Anchor" API in my project. When I try to Cloud Save my anchors, sometimes it succeeds but sometimes it fails. And somecase this Quest 2 device works but the other Quest 2 devices doesn't work. Please tell me what might be the cause of the failure.
FYI, I get the logs below when the cloud save fails:
2023-04-19 10:27:27.178 4616-5972 [CT] com...k.spatial_persistence_service E xr_vega_map_service: Failed to upload & async ingest map with error: std::runtime_error: TigonNetworker: request failure: TransientError domain: CURLcode code: 28 detail: 28: Failed to connect to graph.oculus.com port 443 after 30366 ms: Connection timed out
2023-04-19 10:27:27.179 4616-5972 [CT] com...k.spatial_persistence_service E xr_cloud_anchor_service: Failed to upload spatial anchor with error: std::runtime_error: Failed to upload map
2023-04-19 10:27:27.179 4616-5972 [CT] com...k.spatial_persistence_service I CasNetworkTelemetryObserver: CasNetworkTelemetryObserver onError
2023-04-19 10:27:27.181 5640-5725 [CT] com.stoicdev.ssatestprojectref E SP:AF:AnchorFrameworkSlamAnchor: saveAnchor failed!
19 Replies
I've had this same issue... both with the Cloud save, and the Cloud LoadUnboundAnchors. Sometimes they work, and sometimes they don't. The pattern, for me at least, seems to be that it works the first time... but if I try again immediately, it fails until I reboot the headset or wait an hour or so. If I do the exact same calls to local instead of cloud, everything saves and loads always as expected. One thing that I suspect is that I'm in the public test channel, and maybe its a bug in the beta version. I am on V51 (and it just updated to V53) ... are you, as well? Unfortunately, I can't go back to the release version V50 to test that theory because of the Android 12 update that the PTC did.
Glad that I'm not the only one with the issue.
I'm not in the public test channel, and my device is on V50, so the issue still happens even on v50
Would be nice to get some official confirmation on this one. We're struggling with the same. Just upgraded our app to IntegrationSDK v50 and using a Quest Pro with OSv50 we get basically the same error message as you have posted. To be clear, we did get the same error when on Integration SDK v47. (Which did work a month ago with the exact same build)
Maybe it's a problem with recent Unity versions. I'm on 2022.2.15f1 ... how about you all?
I'm on Unity 2021.3.22f1
Also on 2021.3.22f1
Good news... the new V51 Integration package seems to have fixed this issue for me. 🙂
I updated it today too but the results are being the same, sometimes it works, some it doesn't, and happens both on sharing and loading to and from cloud, did you make any other change to your project so it works consistently?
Same situation as Atrujanox. I'm on V51 but still getting the same result : somtimes cloud save works, but most of the time it doesn't work
Well, that's very interesting. I did implement a new call to "Localize" the unbound anchors upon loading them that I didn't have before - that's the only change since we last updated each other here, but that shouldn't have fixed anything in the save process. But I confirmed again this morning, the whole cloud "create", "save", "share", "load" process is working consistently between all friends and between my QP and Q2.
I am thinking this has something to do with project status, is your project fully verified? I have mine still in draft and I am thinking (but only speculating) that the cloud services are somewhat limited to a few queries each hour
I'm not sure what you mean by "project status" - my app has been uploaded to alpha channel only, and I have never actually submitted anything for review or actual release by Meta.
I am coming to a solution that has something to do when the app is re-initiated, leaving the app stopped some minutes to then try again makes it work well, but if I close the app and then open it again after some seconds looks like the saving and loading behaviours stop responding, not even giving error or success messages.
Maybe the solution is closing a session with oculus services or something before closing the app. Which I don't still know how to do it but will investigate further.
