HSETEX values become "7" when syncing nodes

[kristjgr]

Describe the bug
Hello, we are experiencing a weird bug where all our values become “7” using the HSETEX command when syncing dragonfly nodes.

To Reproduce

We have two 8gb dragonfly nodes in kubernetes and one operator. Version v1.1.9 of Helm Chart for operator, and we are running with backup: dbfilename=anbdb with hourly snapshot.

Populate the database with:
HSETEX a 86400 x 1 y 3 z 4
and let it sync, then if I delete one of the pods, then the values after syncing in the deleted pod will become all “7”:

hgetall a
1) "x"
2) "7"
3) "z"
4) "7"
5) "y"
6) "7"

Expected behavior
Expected values to be as before sync.

Screenshots
NA

Environment (please complete the following information):

• OS: Ubuntu 22.04.5 LTS

• Kernel: 5.15.0-1086-azure

• Containerized?: kubernetes

• Dragonfly Version: 1.26.2

we experienced this half a year ago too (didnt report it though) but all the values were then "6" - so at least now we know its not the devil roaming in there

[adiholden]

Thanks for reporting this @kristjgr

[adiholden]

runing the command
HSETEX a 86400 x 1 y 3 z 4
restart dragonfly that loads the snapshot will result with this bug