Deployment Failures: "Partial Failure" and "flow may be impaired"

[williamking]

I have been unable to successfully deploy to dev (using apigeetool locally) or from dev to test today.

When deploying new revision to dev, the error result contains (among other things):

"Partial Failure"

When deploying from dev to test (in EDGE management portal), the error is:

The revision is deployed and traffic can flow, but flow may be impaired. Unexpected error occurred while processing the updates

(I've seen the second one before and I know there are at least a couple of forum issues related to it.)

I wanted to post both an answer and a question:

SOLUTION: The seemingly sure way to resolve this is to have nothing deployed in the target environment before you perform the new deployment. This obviously is not ideal because there will be no proxy deployed for however long it takes EDGE to deploy the new revision.

QUESTION: Why is this happening? Is there anything I can do to help reduce the chances this occurs?

[#2]

Hi there!

You should be able to do deployments seamlessly without running into these issues. There are a few ways to check what went wrong during deployment. You can check the deployment status using this API: 'https://api.enterprise.apigee.com/v1/o/{orgname}/e/{envname}/apis/{apiproxyname}/deployments'

To achieve zero downtime deployment: curl -X POST -H "Content-type:application/x-www-form-urlencoded" \https://api.enterprise.apigee.com/v1/o/{org_name}/environments/{env-name}/apis/{api_name}/revisions/... \-u username:password

as explained here: http://apigee.com/docs/api-services/content/deploy-api-proxies-using-management-api

Please give it a try and let me know if this helps.

Thanks,

Archendra

[#3]

FYI - you can see that I added the tag "partial-deploy-error" above. I would ask that anyone that sees this problem in the forums, or posts about it add this tag in the future please 🙂

[#4]

FYI - you can see that I added the tag "partial-deploy-error" above. I would ask that anyone that sees this problem in the forums, or posts about it add this tag in the future please 🙂

[williamking]

The apigeetool module is what ultimately performs the deployment. It uses the above parameters as part of its call to the proxies management api.

[remeeshnair]

Hi William,

I have seen sometimes this happens due to connection broken between Router-Message Processor nodes and Management Server (OPDK4.15.04), as first step try redeploying as @archendra suggested above. If it is repeatedly failing check your RMP nodes are up and in cluster.

On Each Router-MP box

Router(port default 8081) : curl -v "http://localhost:port/v1/cluster/members"

MP(port default 8082) : curl -v "http://localhost:port/v1/cluster/members"

If that is fine then try a management server restart might be required followed by a deployment. I am not sure this connection broken retry is something getting fixed in any latest version, which might avoid this MS restart.This is from my expereince.

[#7]

FYI - you can see that I added the tag "partial-deploy-error" above. I would ask that anyone that sees this problem in the forums, or posts about it add this tag in the future please 🙂

[williamking]

I am using public cloud and do not have access to the above functionality

[#9]

if this is still happening you should open a support ticket immediately and raise the same issues. You can even point to this ticket as a point of reference... that said - they might need your ORG name in the support ticket (dont put it here)

[#10]

are you a private cloud or public cloud user?

[williamking]

public cloud

[#12]

FYI - you can see that I added the tag "partial-deploy-error" above. I would ask that anyone that sees this problem in the forums, or posts about it add this tag in the future please 🙂

[#13]

@Birute Awasthi - id love to recruit you on this too....