[Error] - The file cannot be accessed right now

Docs
, ,
1

Hello,
I’m facing the error in OnlyOffice editor: “The file cannot be accessed right now”.
My environment description:

  • Alfresco as a content storage with installed alfresco integration OO module.

  • Document server in docker on AWS ECS:
    – LB to balance requests between 2 ECS tasks

    – ECS service(2 tasks):
    docker image: onlyoffice/documentserver-de:6.4
    ram: 8192
    cpu: 1024

    – Amazon MQ:
    Broker engine: RabbitMQ
    Broker instance type: mq.t3.micro
    Broker engine version 3.8.22
    Deployment mode: Single-instance broker

    – ElastiCache:
    Engine: Redis
    Engine Version Compatibility: 5.0.6
    Node type: cache.t2.micro
    Number of Nodes: 2 nodes

All works well, but sometimes when the user opens the document in OO editor (instantinated with js web API), such error popup occurs:

image
image1924×738 33.6 KB

The tricky thing is that if user request is being balanced to ECS node 1, the OO editor works well, if user gets to ECS node 2, the user sees above error.

The only error that I see in the Node2 (the error one) logs - /var/log/onlyoffice/documentserver/docservice/out.log:
[2022-03-16T21:45:22.628] [WARN] nodeJS - error description: docId = undefined errorId = Server shutdow
[2022-03-16T21:47:03.820] [WARN] nodeJS - error description: docId = undefined errorId = Server shutdow
[2022-03-17T01:55:16.557] [WARN] nodeJS - error description: docId = undefined errorId = Server shutdow
[2022-03-17T02:06:13.707] [WARN] nodeJS - error description: docId = undefined errorId = Server shutdow

Any thoughts on what needs to be checked/monitored/traced.

Thanks in advance.

2

Hello @AlexLebe.dev
Unfortunately, I’m not sure how to help you with your cluster installation. As I see you didn’t use Kubernetes, but your own way. I think the best way for you is to submit a ticket on https://onlyofficesupport.zendesk.com/
I will notify my colleagues, they will try to help you.

3

Thank you for your reply @Alexandre,
I’ve reproduced this issue in my local environment:

  • same storage;
  • dockerized DocumentServer community edition - v. 5.4.2.46

This error occurs when the variable “shutdownFlag = true” in DocsCoServer.js file.

As I understand, this variable is triggered when the server is about to shut down.

I’m curious, how the OO editor (UI component) was initialized, but before downloading the document content (URL provided in editor config), the OO server was not operable.

I suppose the monitoring system did not get that the second node of OO is not working anymore, but LB was still redirecting the requests to that node, but it’s only a guess for now.

4
  • dockerized DocumentServer community edition - v. 5.4.2.46

Oh, this is a really old version of Document server, but I’m glad that you found the reason of the issue and you are continuing investigation.

Anyway, if you need any help with your cluster, it’s better to contact us via Zendesk.

By the way, take a look at Kubernetes guide: https://github.com/ONLYOFFICE/Kubernetes-Docs

I believe, it’s the best cluster solution for Document Server at the moment.

5

Thanks, @Alexandre
for now, I’ll close the thread and check the k8s solution.

1 Like