Signing fails on Github Actions


I’m trying to build my Tizen Wearable app on Github action to automatise my CI like for all other devices I target with my app (Windows, Linux, iOS, Android, …). All the build works fine but I’m not able to sign and got the following output: Signing… keystore password was incorrect

Signing… keystore password was incorrect

However exactly the same calls on my Ubuntu dev machine is passing. I have checked that all the p12 and crt files are correctly available on the CI server. No clue what it can be that make the signing fails.

I found a few threads on the web mentionning this exact same failure but no resolution for that.
Has anyone encounter this same issue ? What would you advise to test to try to understand where the issue is from ?


Best regards,


HI Loic,

I moved this to the Galaxy Watch → Tizen Studio forum as it makes very little sense on GWS forum.

I have a guess but it could be totally wrong. I have GWS and created a Distributor and Author Certficates which are stored in C:\users\USERNAME\GearWatchDesigner\Keystore

My tizen profile uses the same the Author Certificate stored in GearWatchDesigner just fine. My Distributor Certificate is the same place, same password but Tizen Studio Certificate manager gives me an invalid password when I try to use the GWS Created Password. This may be the issue you are seeing.

In my personal computer I uninstalled Tizen Studio and installed Tizen Studio 4.1 and that does not use the Keystore but uses C:/user/USERNAME/SamsungCertificates/ folder with a different distributor password file there.

But I’m only guessing

Samsung Developer Releations

I have had that error while remotely developing on Ubuntu due to Tizen storing certificate passwords as part of the user’s Gnome Keyring.

Even if I am building via CLI I need to have unlocked the keyring which I do by connecting with Remote Desktop, running the browser and entering my password. After that the CLI builds (connected by SSH) will build ok until the machine is rebooted.

No idea if it could be due to a similar issue, but I thought it worth mentioning.