Matt Ober

Would it be possible to support the go-ipfs command even if js-ipfs hasn't implemented the command yet? We plan on utilizing this command for upcoming filecoin support so at least...

@lanzafame I don't have error logs right now, but I can put some together if needed. Admittedly, it concerns me whenever having an application error on purpose becomes part of...

@hsanjuan is describing what I had in mind. Rather than needing to constantly watch an IPFS node to know if it's in a GC, I figured we could actually run...

Another thing to keep in mind here is that we probably want to keep the cluster API itself working on the nodes as they go through GC. This would be...

@lanzafame Good to know. That's an important distinction. So in terms of GC, this would mean that we would need to have the allocator component on each node updated when...

That would be more than fine by me for right now (having cluster manager the GC and wait for the return). @Stebalien do you see anything wrong with that approach?...

@lidel just now seeing this and definitely supporting this change. Count is super expensive for larger pinsets and can cause a lot of issues at scale. I'm in favor of...

@lidel I would also add here that getting rid of count would allow us to raise our rate limits, which is something I know the PL team / ipfs-desktop users...

We could pretty easily implement things right away if we did time based pagination, as we already switched the rest of our application over to doing this, but I'm guessing...

@SgtPooki @lidel did we want to make official the decisions from ipfs-thing?