Postgres clone failed to create volume

Questions / Help
,
1

Hey there, I have a single Postgres machine I’ve been testing Fly out with, and I’m trying to add a single replica. According to the docs I should be able to do that with

fly machines clone <machine-id> --region <region> --app <app>

But when I do that I get the following:

Cloning Machine <machine-id> into region <region>
Volume 'pg_data' will start empty
Error: failed to create volume: Tag is not compatible with the specified digest (Request ID: <some-id>)

Any idea what might be going on here?

2

Hey there,

I just ran that exact command- fly machines clone <machine-id> --region <region> --app <app> and was able to successfully clone a machine.

When running that command are you replacing each flag with what is needed? Maybe there was a typo when you typed it out? For example this was the exact command I ran that was successful:

fly machines clone 5683d49da04368 --region iad --app tipcalculator

3

These are my machines

ID            	NAME             	STATE  	REGION	IMAGE                             	IP ADDRESS                    	VOLUME              	CREATED             	LAST UPDATED        	APP PLATFORM	PROCESS GROUP	SIZE
148eddeefe5498	spring-paper-2404	started	yyz   	flyio/postgres-flex:15.6 (v0.0.49)	fdaa:8:c64a:a7b:f2:a98f:b719:2	vol_450q1yly2y6l018r	2024-03-17T15:22:11Z	2024-03-20T21:40:45Z	v2          	             	shared-cpu-1x:256MB

The exact command I ran was

% fly machines clone 148eddeefe5498 --region yyz --app cohdb-production                                                                            ~/code/cohdb/next fly +
Cloning Machine 148eddeefe5498 into region yyz
Volume 'pg_data' will start empty
Error: failed to create volume: Tag is not compatible with the specified digest (Request ID: 01HSKYFVN4WDESAQGD8BKR71AC-ewr)
4

Hi… I saw this error, as well, when cloning ewr → bos. [It was a single-node, development configuration at first.]

fly image update fixed it for me—and it appears that you, too, are unknowingly running a slightly out-of-date image.

(This might be related to the newly announced image-aware placements.)

Aside: I don’t think having two nodes in the primary region fully works for PG Flex. It’s either 1 or 3, from what I understand.

5

Added volumes

6

This issue should be fixed now. If you do encounter this bug again, please let us know!

1 Like
7

Apologies for the delay, didn’t have time to revisit this until now. I did end up re-deploying my Postgres image and it looks like it bumped the version and resolved the issue. Thanks for the help everyone!

1 Like
8

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.