@kduval I tried creating a pre-build as you suggested but still no luck. I’ve got an error: Unknown deployment error: TypeError: fetch failed.
I sent you the otto logs from the source and the destination servers.
Jim
@kduval I tried creating a pre-build as you suggested but still no luck. I’ve got an error: Unknown deployment error: TypeError: fetch failed.
I sent you the otto logs from the source and the destination servers.
Jim
Hey Jim,
Is there a firewall or something similar on your destination server? It looks like it is having issues sending network requests out. Could you also verify that all of the certs for the server are in place and working: https://www.digicert.com/help/
Something about the network on your server is failing which is causing this issue. The “fetch failed” errors you see in the otto-info log on the destination are trying to make a simple HTTP request out from your server and failing.
-Kyle
@kduval I didn’t found any strange settings in the firewall and I tried running ping and curl commands and it can access external networks just fine.
The SSL for both servers is in place and working.
Today I performed the following test:
So I have three servers:
I created a single test file to copy between those servers. I have 6 combinations of the copy operation in total:
Among the six copy operations, only the MX to SG failed. The other went successfully.
In addition, I found some error messages (please see attachment) in the system log from the SG server using the command: journalctl -u ottofms-proofgeist-com
I don’t have the same error messages from the MX server aside from the message:
Jun 17 08:35:08 fx-mexico systemd[1]: Started ottofms-service-proofgeist.
ottofms-proofgeist-com.zip (5.6 KB)
Hey Jim,
Those errors have a “ConnectTimeoutError” code, which is a sure sign that this is a network problem. If there are no firewall setups or anything else then you may just have shaky connection between the two servers.
I have put together a “fix” in the next version that simply makes us retry the connection a few times when getting information on the build, but if the underlying network issue is not fixed you’ll likely continue having issues.
Unfortunately there is not much other knowledge I can offer outside of “its a network problem”, others in the community are welcome to chime in with their troubleshooting suggestions.
-Kyle
@kduval thank you for keeping with this. So I decided to start from a fresh and newly installed Ubuntu 22.04. I followed the following to install FileMaker Server and OttoFMS:
Install FileMaker Server
Installing FileMaker Server on a single machine (Linux)
Install SSL
ClarisPKB
Install OttoFMS
Installing OttoFMS | OttoFMS
Everything was installed aOK. So I’ve gone ahead and run my first deployment which is a failure (see image).
Would you have another look at this? I’ve sent you the debug info from your Troubleshooting article (Troubleshooting OttoFMS | OttoFMS)
Hey Jim,
OttoFMS version 4.6.0 has hardens this call, so it should try multiple times before failing completely. In your case this likely won’t solve the issue entirely.
I would recommend using the verify server connection endpoint to test that the servers can talk to each other. If that endpoint fails then it confirms that there is a network problem where OttoFMS cannot connect to your other server.
Your error log from this post also confirms for me that this is a network issue. your server is having trouble reaching out and receiving a response.
-Kyle
@kduval I have tried the verify server connection endpoint from my MX to SG server and I got the below response (see also photo):
{"response":{},"messages":[{"code":18000,"text":"Token not found or authentication failed."}]}
It says invalid token or authentication failed. I’m not sure which token or authentication it refers to.
If this is about the Admin API key, I already set one from OttoDeploy for both source (MX) to destination (SG) servers.
Admin console credentials are already in place.
OttoFMS license? I got two of them from the email and I used one each (source/destination). I also tried using only one before and it was working fine.
Root access? I don’t recall this is needed as per the OttoFMS install documentation so maybe it’s not about this.
I asked our network admin and he confirmed that there were no network issues he had found in communicating between the two servers. There were also no additional firewall rules applied.
Would you be available for a quick call (GMT+8) to try and take a look? I’m just asking for a chance here.
Thank You,
Jim
Hey Jim,
Are you available to sign up for Office Hours That’s the best way to get some of our time to take a look.
-Kyle
Thank you for bringing that up. I just made a booking!
Jim
We got through this in a Office Hours session. And it is sort of cool how.
We couldn’t figure out why the deployment was unable to fetch the files from the source server. We’ll look into trying to get more information about this error.
BUT because we just shipped Builds that can push to a server. We just flipped it around.
Instead of Fetching the files from the source server after the build, We had the source server push the files back to the destination after the build was done.
Worked great! Problem solved.
The option to PUSH a build instead of build it is in OttoDeploy. Make sure you have advanced options on, and then look at the Build Options.
YAY
Todd
@toddgeist
Thank you so much for your help during the office hour session last month—it was really insightful, and I appreciate the time taken to assist me.
However, I’ve encountered the issue again while using OttoFMS version 4.6.2. The error details are as follows:
Phase: unknown phase
Message: Unknown deployment error: TypeError: fetch failed
To troubleshoot, I’ve tried the following steps:
While these steps have occasionally worked, it’s been inconsistent—sometimes the deployment succeeds, but it’s a bit hit or miss.
I noticed that version 4.7 has been released. Has this issue been addressed in the latest version? I’ll try updating and see if that helps resolve it, but I’d appreciate any insights you might have.
Jim
Hi Jim,
Are you doing what we discussed in the Office hours session? Are you pushing the build to the other server?
If you can send us the logs for the deployment we could look in more detail. Here is how you can get those.
Todd
@toddgeist, yes, I did those. First, I tried turning on the push and it failed. Then I tried turning it off next. That failed also. I sent you two failed deployment logs from two different files. I performed both this morning.
Jim
Hi Jim
Your logs are reporting that your build server is unreachable. That means that the two servers can’t connect. There appears to be a network issue between them. Did you try the test we discussed last time?
https://docs.ottofms.com/apidoc#tag/Miscellaneous/operation/GetOttoApiCanReachOtherServer
Thanks
Todd
@toddgeist
Yes, I tried querying one server to the other and vice versa using Postman. Both returned success (200 OK).
Jim
Hi Jim,
Ok, that is too bad, since it makes this more mysterious. I am afraid it just isn’t clear why your servers are having difficulty maintaining a connection. But the logs are pretty clear when one server tries to connect to the other server to trigger a build it isn’t able to.
You could try upgrading to OttoFMS 4.7.0. It’s worth trying. But I am not sure if it will help.
You could try sigining up for another office hours. We can take a look again.
Todd