panic: runtime error: index out of range [0] with length 0

[yarongilor]

panic: runtime error: index out of range [0] with length 0

[roydahan]

I think the issue is already fixed.
Yaron, which Gemini version was used?

@fruch / @dkropachev are we using the latest Gemini docker in SCT?
If so, we probably need to backport it as well.

[fruch]

I think the issue is already fixed. Yaron, which Gemini version was used?

@fruch / @dkropachev are we using the latest Gemini docker in SCT? If so, we probably need to backport it as well.

there a PR with latest tag from gemini:
scylladb/scylla-cluster-tests#6115

but from a few runs I've did it doesn't seem to be more stable, if you want to push and backport blindly like that, be my guest.

[yarongilor]

'gemini_version': '1.7.8'

[dkropachev]

@fruch , @roydahan , this particular issue is fixed in master, let me see if we can release it

[dkropachev]

but from a few runs I've did it doesn't seem to be more stable, if you want to push and backport blindly like that, be my guest.

Can I have them ?

[dkropachev]

@fruch , @soyacz , could you please guys release v1.8.1 ? If you have docker creds and aws creds it is enought go run goreleaser after you checkout v1.8.1

[dkropachev]

this particular issue is not gocql-related, it comes from the fact that gemini was manipulating with query values at the same time when it was feeding these values to the gocql.
But there are other gocql bugs that causes panics on serialization/deserialization, I have created PR to address them on the mainstream, once it is merged we can issue new version of scylladb/gocql to merge them in and use it here in gemini.

[roydahan]

@dkropachev shouldn't we also bump the gocql driver version we use in Gemini.
Isn't it something very old?

[dkropachev]

@roy, on the master we have it most modern version - v1.7.3