scylla-cluster-tests icon indicating copy to clipboard operation
scylla-cluster-tests copied to clipboard

set the scylla monitoring image to 4.8.0

Open amnonh opened this issue 1 year ago • 2 comments

Set the monitoring image to version 4.8.0

amnonh avatar Jul 15 '24 13:07 amnonh

@amnonh

The image in AWS isn't found on eu-west-1, is those images synced across regions ?

fruch avatar Jul 28 '24 16:07 fruch

They should be, CC @yaronkaikov, who is responsible for image creation.

amnonh avatar Jul 28 '24 16:07 amnonh

@amnonh

The image in AWS isn't found on eu-west-1, is those images synced across regions?

It's there

image

yaronkaikov avatar Aug 04 '24 08:08 yaronkaikov

@amnonh The image in AWS isn't found on eu-west-1, is those images synced across regions?

It's there

image

@yaronkaikov, we need the Name tag, so we can find it.

fruch avatar Aug 04 '24 16:08 fruch

@amnonh The image in AWS isn't found on eu-west-1, is those images synced across regions?

It's there image

@yaronkaikov, we need the Name tag, so we can find it.

There is one

image

yaronkaikov avatar Aug 04 '24 16:08 yaronkaikov

@amnonh The image in AWS isn't found on eu-west-1, is those images synced across regions?

It's there image

@yaronkaikov, we need the Name tag, so we can find it.

There is one

image

o.k. the code in SCT is looking up by name, only look in one account, and not in the images account. that part was used for fetching other images, and not scylla images by name

fruch avatar Aug 04 '24 16:08 fruch

@yaronkaikov

a PR to fix it, and lookup on both accounts, so monitor image can be found: https://github.com/scylladb/scylla-cluster-tests/pull/8232

previously the image were clone into the main account (since they were available on all regions) we should stop doing that..., since they now should be copied into all regions to begin with.

fruch avatar Aug 04 '24 17:08 fruch

I've recreated 4.8.0 this morning, we should re-generate the images

amnonh avatar Aug 04 '24 17:08 amnonh

I've recreated 4.8.0 this morning, we should re-generate the images

Why re-creating? now we would have two images which are 4.8.0,

yaronkaikov avatar Aug 04 '24 17:08 yaronkaikov

@yaronkaikov 4.8.0 is in RC mode, and I made changes, so I had to recreate the release, I'm not sure if images were created.

amnonh avatar Aug 04 '24 18:08 amnonh

@yaronkaikov 4.8.0 is in RC mode, and I made changes, so I had to recreate the release, I'm not sure if images were created.

@amnonh So let's call it (tag name) 4.8.0-rcX, and then you can increase the number for every change until it's officially out. This way QA for example can distinguish between each image by name

yaronkaikov avatar Aug 04 '24 18:08 yaronkaikov

@yaronkaikov image names contain the date, I can add an rcX to the tags, if it's supported

amnonh avatar Aug 04 '24 18:08 amnonh

@yaronkaikov image names contain the date, I can add an rcX to the tags, if it's supported

Thanks, it should be ok, as we do in Scylla image

yaronkaikov avatar Aug 05 '24 04:08 yaronkaikov

I've found one more issue with the fact the images are on the releng account, and that they are not public https://github.com/scylladb/scylla-cluster-tests/pull/8258

one that would get reviewed and merged, we can do one more round

@amnonh did you created new images ?

fruch avatar Aug 06 '24 15:08 fruch

I did: AWS: scylladb-monitor-4-8-0-2024-08-06t03-34-43z GCP: scylladb-monitor-4-8-0-2024-08-06t03-34-43z

amnonh avatar Aug 06 '24 15:08 amnonh

I did: AWS: scylladb-monitor-4-8-0-2024-08-06t03-34-43z GCP: scylladb-monitor-4-8-0-2024-08-06t03-34-43z

@amnonh @yaronkaikov those images aren't working...

botocore.exceptions.ClientError: An error occurred (AuthFailure) when calling the RequestSpotInstances operation: Not authorized for images: [ami-02f241d7e29ae98ad]

fruch avatar Aug 07 '24 10:08 fruch

I did: AWS: scylladb-monitor-4-8-0-2024-08-06t03-34-43z GCP: scylladb-monitor-4-8-0-2024-08-06t03-34-43z

@amnonh @yaronkaikov those images aren't working...

botocore.exceptions.ClientError: An error occurred (AuthFailure) when calling the RequestSpotInstances operation: Not authorized for images: [ami-02f241d7e29ae98ad]

meanwhile I manually setup the sharing, and raised: https://github.com/scylladb/scylla-monitoring/issues/2375

for the next time we need those images

fruch avatar Aug 07 '24 14:08 fruch

@amnonh @fruch I suppose this fix broke some Manager tests:

13:01:34  ----- LAST ERROR EVENT -------------------------------------------------------
13:01:34  2024-08-13 11:01:20.621: (TestFrameworkEvent Severity.ERROR) period_type=one-time event_id=8e220cb6-4ea4-4484-9223-5099475f54c2, source=MgmtCliTest.SetUp()
13:01:34  exception=An error occurred (AuthFailure) when calling the RequestSpotInstances operation: Not authorized for images: [ami-0b02bf09acc3d3292]

https://jenkins.scylladb.com/view/scylla-manager/job/manager-master/job/ubuntu22-installation-test/2/

mikliapko avatar Aug 13 '24 11:08 mikliapko

@amnonh @fruch I suppose this fix broke some Manager tests:

13:01:34  ----- LAST ERROR EVENT -------------------------------------------------------
13:01:34  2024-08-13 11:01:20.621: (TestFrameworkEvent Severity.ERROR) period_type=one-time event_id=8e220cb6-4ea4-4484-9223-5099475f54c2, source=MgmtCliTest.SetUp()
13:01:34  exception=An error occurred (AuthFailure) when calling the RequestSpotInstances operation: Not authorized for images: [ami-0b02bf09acc3d3292]

https://jenkins.scylladb.com/view/scylla-manager/job/manager-master/job/ubuntu22-installation-test/2/

In which region is this running?

I can guess I've missed some AMI, setting the sharing information

Until https://github.com/scylladb/scylla-monitoring/issues/2375 Would be fixed

fruch avatar Aug 13 '24 12:08 fruch

@amnonh @fruch I suppose this fix broke some Manager tests:

13:01:34  ----- LAST ERROR EVENT -------------------------------------------------------
13:01:34  2024-08-13 11:01:20.621: (TestFrameworkEvent Severity.ERROR) period_type=one-time event_id=8e220cb6-4ea4-4484-9223-5099475f54c2, source=MgmtCliTest.SetUp()
13:01:34  exception=An error occurred (AuthFailure) when calling the RequestSpotInstances operation: Not authorized for images: [ami-0b02bf09acc3d3292]

https://jenkins.scylladb.com/view/scylla-manager/job/manager-master/job/ubuntu22-installation-test/2/

In which region is this running?

I can guess I've missed some AMI, setting the sharing information

Until scylladb/scylla-monitoring#2375 Would be fixed

It's us-east-1 region

mikliapko avatar Aug 13 '24 12:08 mikliapko

@amnonh @fruch I suppose this fix broke some Manager tests:

13:01:34  ----- LAST ERROR EVENT -------------------------------------------------------
13:01:34  2024-08-13 11:01:20.621: (TestFrameworkEvent Severity.ERROR) period_type=one-time event_id=8e220cb6-4ea4-4484-9223-5099475f54c2, source=MgmtCliTest.SetUp()
13:01:34  exception=An error occurred (AuthFailure) when calling the RequestSpotInstances operation: Not authorized for images: [ami-0b02bf09acc3d3292]

https://jenkins.scylladb.com/view/scylla-manager/job/manager-master/job/ubuntu22-installation-test/2/

In which region is this running? I can guess I've missed some AMI, setting the sharing information Until scylladb/scylla-monitoring#2375 Would be fixed

It's us-east-1 region

Fix the permissions, and re-run that job

fruch avatar Aug 13 '24 14:08 fruch