Oran Agra

sorry for the luck of update, an update was posted in the above referenced #8180. and also just posted it in #7382 (and closed it). in essence, we concluded that...

@Super-long it took me some time to study that area since i didn't dig too much into it in the past. on the surface it looks like a logical change,...

i see there's already a PR for that: #10874 (at least the ZSET part). maybe you can help evaluate it.

I'm not familiar with some of these details myself either.. i'll try to grab the attention of someone who does.. but also, i guess that for the case of redis.io...

@ThachNgocTran i think the fact you use `redis://` after the `--tls` means it overrides the TLS flag. you should either use `rediss://` or put the `--tls` after the `-u`.

@itamarhaber is that a bug you intend to fix or ask someone else to fix?

We had a discussion about this with Yossi, we concluded that it's better / safer to let the module that saved this field to declare if loading it is optional...

the difference is that for keys, the user can easily delete the module keys before saving the rdb. for aux it's not that simple, the module can't easily avoid saving...

the problem i wanna solve is a case were someone has an instance with a module which he didn't use, or use it and then stopped using it, and now...

the problem with the current approach in this PR is that it could be that some modules save their entire data in the aux field, and silently skipping it could...