I've configured upgrades on my VSTS release of a Service Fabric app containing 5 services to a single node test environment on Azure. Unfortunately when it gets to the release part it just hangs saying "Waiting for upgrade..." over and over again. I left it for 15 hours and it still says the same thing. The initial deployment went ahead without issue.
I've looked at various posts about turning off health check times, but this has not been successful. I've also tried setting the mode to UnmonitoredAuto, but no success.
I've RDPd onto the environment and checked the processor/memory usage in task manager, and everything is pretty much 0%, and very low memory usage.
Is there anything else I can do to stop the upgrade hanging?
OK, I've managed to fix this. This was happening because there is a PreUpgradeSafetyCheck that happens before rolling out an upgrade. This is not relevant for a single node cluster as downtime is inevitable for single node clusters.
The status of an upgrade can be found using: Get-ServiceFabricApplicationUpgrade. Which shows the status above.
To fix this there is a flag: UpgradeReplicaSetCheckTimeoutSec in the release task. Setting the value to 0 sorts things out.
Related
I have an ECS cluster (fargate), task, and service I have had setup in Terraform for at least a year. I haven't touched it for a long while. My normal deployment for updating the code is to push a new container to the registry and then stop all tasks on the cluster with a script. Today, my service did not run a new task in response to that task being stopped. It's desired count is fixed at so it should.
I have go in an tried to manually run this and I'm seeing this error.
Unable to run task
Http request timed out enforced after 4999ms
When I try to do this, a new stopped task is added to my stopped tasks lists. When I look into that task the stopped reason is "Deployment restart" and two of them are now showing "Task provisioning failed." which I think might be tasks the service tried to start. But these tasks do not show a started timestamp. The ones I start in the console have a started timestamp.
My site is now down and I can't get it back up. Does anyone know of a way to debug this? Is AWS ECS experiencing problems right now? I checked the health monitors and I see no issues.
This was an AWS outage affecting Fargate in us-east-1. It's fixed now.
Firstly, yes I have read this https://www.liquibase.com/blog/using-liquibase-in-kubernetes
and I also read many SO threads where people are answering "I solved the issue by using init-container"
I understand that for most people this might have fixed the issue because the reason their pods were going down was because the migration was taking too long and k8s probes killed the pods.
But what about when a new deployment is applied and the previous deployment was stuck a failed state (k8s trying again and again to launches the pods without success) ?
When this new deployment is applied it will simply whip / replace all the failing pods and if this happens while Liquibase aquired the lock the pods (and its init containers) are killed and the DB will be left in a locked state requiring manual intervention.
Unless I missed something with k8s's init-container, using them doesn't really solve the issue described above right?
Is that the only solution currently available? What other solution could be used to avoid manual intervention ?
My first thought was to add some kind of custom code (either directly in the app before the Liquibase migration happens) or in init-container that would run before liquibase init-container runs to automatically unlock the DB if for example the lock is, let's say, 5 minutes old.
Would that be acceptable or will it cause other issues i'm not thinking about ?
This morning we noticed that all Kubernetes clusters in all projects ( 2 projects, 2 clusters per project ) showed unavailable / ERROR in the Google Cloud Console.
The dashboard shows no current issues: https://status.cloud.google.com/
It basically looks like the master nodes are down, the API does not respond and the clusters cannot be edited in the UI. Before the weekend everything was up and since at least yesterday evening they all show in red.
The deployed services fortunately respond, but we cannot manage the cluster in any way.
I reported it here too:
https://issuetracker.google.com/issues/172841082
Did anyone else encounter this and is there any way to restart or trigger the master node to restart? I cannot edit the cluster so an upgrade is not possible either.
I read elsewhere that only SRE folks from Google can (re)start them.
It's beyond me how this can happen.
By the way, auto-repair is set to on and I followed the troubleshooting page, basically with all paths leading to : master node down, nothing to be done.
Any help would be greatly appreciated, or simply a SRE doing a start node action ;).
Thank you #dany L, it was indeed a billing issue.
I'm surprised there is nothing like a message in the Cloud Console and one has to go to billing specifically to find out about this.
After billing was fixed, it took a few minutes while before the clusters were available, then everything looked back to normal.
When I did the upgrade of concourse from 3.4.0 to 3.5.0, suddenly all running jobs changed their state from running to errored. I can see the string 'no workers' appearing at the start of their log now. Starting the jobs manually or triggered by the next changes didn't have any problem.
The upgrade of concourse itself was successful.
I was watching what bosh did at the time and I saw this change of job states took place all at once while either the web or the db VM was upgraded (I don't know which one). I am pretty sure that the worker VMs were not touched yet by bosh.
Is there a way to avoid this behavior?
We have one db, one web VM and six workers.
With only one web VM it's possible that it was out of service for long enough that all workers expired. Workers continuously heartbeat and if they miss two heartbeats (which takes 1 minute by default) they'll stall. They should come back after the deploy is finished but if scheduling happened before they heartbeats, that would cause those errors.
I Came by a problem where i have an Ops Manager that suppose to run a MongoDB cluster as an automated cluster.
Suddenly the servers started going down, unexpectedly - while there are no errors in any of the log files indicating on when is the problem.
The Ops Manager gets stuck on the blue label
We are deploying your changes. This might take a few minutes
And it just never goes away.
Because this environment is based on the automation feature, the mms is managing the user on the servers and runs all of the processes from "mongod" which i can't access even as a Root (administrator).
As far as the Ops Manager goes it shows that a shard in a replica set is down although it's live, and thinks that a mongos that is dead is alive.
Has someone got into this situation before and may be able to help ?
Thanks,
Eliran.
Problem found: there was an ntp mismatch between the servers in the cluster somehow, so what happened was that the servers were not synced and everytime the ops manager did something it got responses with wrong times and could not use it's time limits.
After re-configuring all the ntp's back to the same one - everything got back to how it should have been :)