Globaleaks in Cobia as Custom App

[lucapsg]

Relieved by @sfatula 's statement I'm trying to install Globaleaks in TNS as Custom App but I can't.

I created the image in Docker Desktop for Windows using the official Dockerfile and docker-compose.yml .
Since it worked the first time, I published it in my personal repo, so I could use it in TNS.

In TNS:
1. I downloaded the image from Apps -> Settings -> Manage Container Images -> Pull Image because the personal repo requires authentication and I see it among those present in the host
2. I created the dataset /mnt/tank1/appdata/globaleaks for its data, specifying Share Type = Apps, also for the parent dataset
3. I started the Custom App wizard as follows:

Application Name: globaleaks1
Image repository: myrepo/globaleaks
Image Tag: latest
Image Pull Policy: Only pull image if not present on host
Container CMD: /usr/bin/python3
Container Args: /usr/bin/globaleaks
Container Args: --working-path=/var/globaleaks/
Container Args: -n
Host Path: /mnt/tank1/appdata/globaleaks
Mount Path: /var/globaleaks

Although I don't think the problem depends on the network settings, I tried using both the Networking section, specifying the same interface used by the host...:

Host Interface: br1
IPAM Type: Use DHCP

...and the Port Forwarding section...:

Container Port: 8080
Node Port: 9080
Protocol: TCP
Container Port: 8443
Node Port: 9443
Protocol: TCP

...leaving the other items at their default values.

Whatever you do, the result is always the same:
- the application is not displayed among those installed (0 Installed Apps), so at each attempt you have to reconfigure everything from scratch
- /var/log/k3s_daemon.log is constantly flooded with messages (this excerpt starts from the moment I saved the wizard settings):

Dec 3 18:01:40 HV04 k3s[6496]: I1203 18:01:40.655343 6496 alloc.go:327] "allocated clusterIPs" service="ix-globaleaks1/globaleaks1-ix-chart" clusterIPs=map[IPv4:172.17.115.169]
Dec 3 18:01:40 HV04 k3s[6496]: I1203 18:01:40.674330 6496 event.go:294] "Event occurred" object="ix-globaleaks1/globaleaks1-ix-chart" fieldPath="" kind="Deployment" apiVersion="apps/v1" type="Normal" reason="ScalingReplicaSet" message="Scaled up replica set globaleaks1-ix-chart-7f9684d74 to 1"
Dec 3 18:01:40 HV04 k3s[6496]: I1203 18:01:40.726854 6496 event.go:294] "Event occurred" object="ix-globaleaks1/globaleaks1-ix-chart-7f9684d74" fieldPath="" kind="ReplicaSet" apiVersion="apps/v1" type="Normal" reason="SuccessfulCreate" message="Created pod: globaleaks1-ix-chart-7f9684d74-4f9mp"
Dec 3 18:01:40 HV04 k3s[6496]: I1203 18:01:40.742814 6496 topology_manager.go:210] "Topology Admit Handler"
Dec 3 18:01:40 HV04 k3s[6496]: I1203 18:01:40.788618 6496 event.go:294] "Event occurred" object="ix-globaleaks1/globaleaks1-ix-chart-7f9684d74" fieldPath="" kind="ReplicaSet" apiVersion="apps/v1" type="Normal" reason="SuccessfulCreate" message="Created pod: globaleaks1-ix-chart-7f9684d74-nxf2l"
Dec 3 18:01:40 HV04 k3s[6496]: I1203 18:01:40.805995 6496 topology_manager.go:210] "Topology Admit Handler"
Dec 3 18:01:40 HV04 k3s[6496]: I1203 18:01:40.858747 6496 event.go:294] "Event occurred" object="ix-globaleaks1/globaleaks1-ix-chart-7f9684d74" fieldPath="" kind="ReplicaSet" apiVersion="apps/v1" type="Normal" reason="SuccessfulCreate" message="Created pod: globaleaks1-ix-chart-7f9684d74-fznv5"
Dec 3 18:01:40 HV04 k3s[6496]: I1203 18:01:40.873532 6496 topology_manager.go:210] "Topology Admit Handler"
... 26 pairs of lines similar to the last two...
Dec 3 18:01:45 HV04 k3s[6496]: I1203 18:01:45.330561 6496 gc_controller.go:207] "Garbage collecting pods" numPods=10
Dec 3 18:01:45 HV04 k3s[6496]: I1203 18:01:45.330619 6496 gc_controller.go:329] "PodGC is force deleting Pod" pod="ix-globaleaks1/globaleaks1-ix-chart-7f9684d74-ft9sd"
Dec 3 18:01:45 HV04 k3s[6496]: I1203 18:01:45.330671 6496 gc_controller.go:329] "PodGC is force deleting Pod" pod="ix-globaleaks1/globaleaks1-ix-chart-7f9684d74-scxph"
Dec 3 18:01:45 HV04 k3s[6496]: I1203 18:01:45.331011 6496 gc_controller.go:329] "PodGC is force deleting Pod" pod="ix-globaleaks1/globaleaks1-ix-chart-7f9684d74-4f9mp"
...7 more lines similar to the last one...
Dec 3 18:01:45 HV04 k3s[6496]: I1203 18:01:45.370994 6496 pod_startup_latency_tracker.go:102] "Observed pod startup duration" pod="ix-globaleaks1/globaleaks1-ix-chart-7f9684d74-gn8lp" podStartSLOduration=1.370961677 pod.CreationTimestamp="2023-12-03 18:01:44 +0100 CET" firstStartedPulling="0001-01-01 00:00:00 +0000 UTC" lastFinishedPulling="0001-01-01 00:00:00 +0000 UTC" observedRunningTime="2023-12-03 18:01:45.359309167 +0100 CET m=+1619.257132725" watchObservedRunningTime="2023-12-03 18:01:45.370961677 +0100 CET m=+1619.268785214"
...

It seems that the application starts but ends immediately after and then K3s restarts it endlessly.

I'm new to Kubernetes and Custom Apps.
What am I leaving out?

 

[lucapsg]

I tried again after installing another app and this time (???) Globaleaks appeared among the installed apps, although it still doesn't work.

In the History box there are many messages like these:

2023-12-05 17:13:19 Allocate failed due to no healthy devices present; cannot allocate unhealthy devices amd.com/gpu, which is unexpected
2023-12-05 17:13:19 Successfully assigned ix-globaleaks1/globaleaks1-ix-chart-57c867dc99-gskgd to ix-truenas
2023-12-05 17:13:19 Scaled down replica set globaleaks1-ix-chart-57c867dc99 to 0 from 1
2023-12-05 17:13:18 Allocate failed due to no healthy devices present; cannot allocate unhealthy devices amd.com/gpu, which is unexpected
2023-12-05 17:13:18 Successfully assigned ix-globaleaks1/globaleaks1-ix-chart-57c867dc99-cr465 to ix-truenas
2023-12-05 17:13:18 Allocate failed due to no healthy devices present; cannot allocate unhealthy devices amd.com/gpu, which is unexpected
2023-12-05 17:13:18 Successfully assigned ix-globaleaks1/globaleaks1-ix-chart-57c867dc99-lpq5x to ix-truenas

It seems that when the application starts the system tries to use hardware not present in the host (cannot allocate unhealthy devices amd.com/gpu), an old Fujitsu PRIMERGY TX100 S3P with Xeon E3-1220 V2 and 24GB of RAM without video card.

Editing the app config I saw that in all three drop-down boxes in the "GPU Configuration" section there is a single item "Allocate 0 xxxx GPU", which would seem to mean that NO GPU resources will be allocated for the application.
The application certainly doesn't require them.

If my deductions are correct why does TNS try to allocate GPU resources?
How can I solve/work around the problem?