Unknown deployment error: Error: Local Admin API error: fetch failed: SocketError: other side closed

In OttoFMS 4.2.2 I get an unknown deployment error on FMS 20.3.1.31.

Unknown deployment error: Error: Local Admin API error: fetch failed: SocketError: other side closed

It’s the first sub migration of a two stage, live copy to testing, then new build migration to testing deployment. Any ideas on what I can do to work round this?

Deployment JSON looks like this (some data redacted).

{
  "source": {
    "type": "otto4jit",
    "serverUrl": "XXX",
    "auth": {
      "type": "apiKey",
      "apiKey": "XXX"
    },
    "buildOptions": {
      "compressionLevel": 0,
      "memoryLevel": 1
    }
  },
  "name": "XXX Live to XXX Testing",
  "fileOperations": [
    {
      "operation": "replace",
      "source": {
        "fileName": "XXX.fmp12",
        "encryptionKey": "XXX"
      },
      "target": {
        "fileName": "XXX - Testing.fmp12",
        "encryptionKey": "XXX"
      }
    }
  ],
  "concurrency": 1,
  "batchId": 2,
  "status": "queued",
  "batchName": "XXX Testing Update"
}

Hi Gabe,

Thanks for reporting this. We are looking into another instance of this error or a similar one. What operating system is running on the servers?

Thanks

Todd

Thank’s Todd. The server is running Ubuntu 22.04 LTS. I have a number of other servers with the exact same setup, same OS version, same FMS version, same Otto version where this does not happen, so I’m not sure what’s different in this instance.

We suspect a problem with FileMaker Server Admin API. It is the “other side” that closed here.

Take a look in this file

/opt/FileMaker/FileMaker Server/Admin/FAC/logs/fac.log

On the machines that are working and the one that isn’t.

I think you will see some differences there. Send us the files if you wouldn’t mind. You can send them via a Direct message if you like. We are gathering data for that might be useful to hand off to Claris in a bug report.

We are also looking into work-a-rounds. I’ll let you know as soon as we have version you can test.

Thanks

Todd

One more thing you could test. If it is a problem with the Admin API, restarting the AdminServer might help.

on the command line fmsadmin restart adminerver -y

If decide to try that and run your deployment again, please let us know what happens.

Thanks

Todd

Restarting the Admin API from the command line give me:

Error: 10007 (Requested object does not exist)

So maybe it’s not running? It’s on a production server so I can’t try restarting it at the moment, but I’ll have a look once I get a chance.

oops typo

fmsadmin restart adminserver -y

Sorry

Todd

Ahh, yeah, didn’t spot that. Yeah same result of the deployment after restarting the admin api.

anything unusual in fac.log?

Also… we think that you can force the Admin API and Admin Console to crash by clicking around in the Admin Console UI on machines where this is a problem.

Is there any chance that you have the Admin Console open on that machine? And are clicking around?

Thanks for looking into this

Todd

Tried again after logging out of the FMS console and still fails. Looking at the fac.log has these errors.

2024-03-07 17:13:37 +0000	Execute /opt/FileMaker/FileMaker Server/Admin/FAC/facstart.sh
2024-03-07 17:13:45 +0000 - error:   fmi   127.0.0.1   adminapi   general   n/a   "[{"code":"1703","text":"Invalid username or password, or JSON Web Token"}]"
2024-03-07 17:13:46 +0000 - error:   fmi   127.0.0.1   adminapi   general   n/a   "[{"code":"1703","text":"Invalid username or password, or JSON Web Token"}]"
2024-03-07 17:13:48 +0000 - error:   fmi   127.0.0.1   adminapi   general   n/a   "[{"code":"1703","text":"Invalid username or password, or JSON Web Token"}]"
2024-03-07 17:13:52 +0000 - error:   fmi   127.0.0.1   adminapi   general   n/a   "[{"code":"1703","text":"Invalid username or password, or JSON Web Token"}]"
2024-03-07 17:14:07 +0000 - error:   fmi   127.0.0.1   adminapi   general   n/a   "[{"code":"1703","text":"Invalid username or password, or JSON Web Token"}]"
2024-03-07 17:16:51 +0000	Stopping FAC node
2024-03-07 17:16:51 +0000	Killing: 453886

So Otto has access and it can close the file, but the copying the just in time build build fails?

I am looking into your logs. Thanks for sending them.

I am working on a test version which I hope will solve this or at least get us more information.

I’ll let you know when it is ready.

Todd

1 Like