I have long running workers running in kubernetes - more than 5 hours. I want to update the container without interrupting the long running jobs. I want any newly started work off the queue to start with the new version of the release but I don't want to interrupt the currently running work.
BTW I'm not actually using Jobs, I'm using Deployments with workers that get work off a redis queue.
What is the best way to do to do a release without killing the long running work?
Have a huge timeout for SIGTERM
preStop hooks?
Another container in the pod that checks for the latest version and updates once work is done?
Related
I am running an airflow cluster on EKS on AWS. I have setup some scaling config for worker setup. If CPU/Mem > 70% then airflow spins up new worker pod. However I am facing an issue when these worker pods are scaling down. When worker pods start scaling down, two things happen:
If no tasks is running on a worker pod, it terminates within 40sec.
If any task is running on a worker pod, it terminates in about 8min, and after one more minute, I find the task failing on UI.
I have setup below two properties in helm chart for worker pod termiantion.
celery:
## if celery worker Pods are gracefully terminated
## - consider defining a `workers.podDisruptionBudget` to prevent there not being
## enough available workers during graceful termination waiting periods
##
## graceful termination process:
## 1. prevent worker accepting new tasks
## 2. wait AT MOST `workers.celery.gracefullTerminationPeriod` for tasks to finish
## 3. send SIGTERM to worker
## 4. wait AT MOST `workers.terminationPeriod` for kill to finish
## 5. send SIGKILL to worker
##
gracefullTermination: true
## how many seconds to wait for tasks to finish before SIGTERM of the celery worker
##
gracefullTerminationPeriod: 180
## how many seconds to wait after SIGTERM before SIGKILL of the celery worker
## - [WARNING] tasks that are still running during SIGKILL will be orphaned, this is important
## to understand with KubernetesPodOperator(), as Pods may continue running
##
terminationPeriod: 120
I can see that worker pod should shutdown after 5 mins or irrespective task running or not. So I am not sure why I see total of 8 min for worker pod termination. And my main issue is there any way I can setup config so that worker pod only terminates when task running on it finishes execution. Since tasks in my dags can run anywhere between few minutes to few hours so I don't want to put a large value for gracefullTerminationPeriod. I Would appreciate any solution around this.
Some more info: Generally the long running task is a python operator which runs either a presto sql query or Databricks job via Prestohook or DatabricksOperator respectively. And I don't want these to recivie SIGTERM before they complete their execution on worker pod scaling down.
This is not possible due to limitations from K8 end. More details are available here. However by using a large value of "gracefulTerminationPeriod" works, although this is not what I intended to do but it works better than I originally thought. When large value of gracefulTerminationPeriod is set, workers doesn't wait around for gracefulTerminationPeriod time to terminate. If a worker pod is marked for termination it terminates as soon as tasks running on it reaches zero.
Until K8 accept proposed changes and new community helm chart is released, I think this is the best solution without incurring costs of keeping worker up.
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 have a working kubernetes cluster (v1.4.6) with an active job that has a single failing pod (e.g. it is constantly restarted) - this is a test, the job should never reach completion.
If I restart the same cluster (e.g. reboot the node), the job is properly re-scheduled and continues to be restarted
If I upgrade the cluster to v1.5.3, then the job is marked as completed once the cluster is up. The upgrade is basically the same as restart - both use the same etcd cluster.
Is this the expected behavior when going to v1.5.x? If not, what can be done to have the job continue running?
I should provide a little background on my problem - the job is to ultimately become a driver in the update process and it is important to have it running (even in face of cluster restarts) until it achieves a certain goal. Is this possible using a job?
In v1.5.0 extensions/v1beta1.Jobs was deprecated in favor of batch/v1.Job, so simply upgrading the cluster without updating the job definition is expected to cause side effects.
See the Kubernetes CHANGELOG for a complete list of changes and deprecations in v1.5.0.
I have a Kubernetes cluster running Django, Celery, RabbitMq and Celery Beat. I have several periodic tasks spaced out throughout the day (so as to keep server load down). There are only a few hours when no tasks are running, and I want to limit my rolling-updates to those times, without having to track it manually. So I'm looking for a solution that will allow me to fire off a script or task of some sort that will monitor the Celery server, and trigger a rolling update once there's a window in which no tasks are actively running. There are two possible ways I thought of doing this, but I'm not sure which is best, nor how to implement either one.
Run a script (bash or otherwise) that checks up on the Celery server every few minutes, and initiates the rolling-update if the server is inactive
Increment the celery app name before each update (in the Beat run command, the Celery run command, and in the celery.py config file), create a new Celery pod, rolling-update the Beat pod, and then delete the old Celery 12 hours later (a reasonable time span for all running tasks to finish)
Any thoughts would be greatly appreciated.
My application uses Sidekiq to handle long (several minutes) running background tasks. Deployments are done with Capistrano 2 and all processes are monitored with Monit.
I have used capistrano-sidekiq to manage the sidekiq process during deployments but it has not worked perfectly. Some times during the deployment a new sidekiq process is started but the old one is not killed. I believe this happens because capistrano-sidekiq is not operating through Monit during the deployment.
Second problem is that because my background tasks can take several minutes to complete my deployment should allow two sidekiq processes to co-exisit. The old sidekiq process should be allowed to complete the tasks it is processing and a new sidekiq process should start taking new tasks into processing.
I have been thinking about something like this into my deploy script
When deployment starts:
I tell Monit to unmonitor the sidekiq process
I stop the current sidekiq process and give it 10 minutes to finish its tasks
After the code has been updated:
I start a new sidekiq process and tell Monit to start monitoring it.
I may need to move the sidekiq process pid file into the release directory if the pid file is not removed until the stopped sidekiq process has eventually been killed.
How does this sound? Any caveats spotted?
EDIT:
Found a good thread about this same issue.
http://librelist.com/browser//sidekiq/2014/6/5/rollback-signal-after-usr1/#f6898deccb46801950f40ad22e75471d
Seems reasonable to me. The only possible issue is losing track of the old Sidekiq's PID but you should be able to use ps and grep for "stopping" to find old Sidekiqs.