Vladislav Zolotarov

@raphaelsc FYI

> doesn't nodetool compactionstats provide some info `compactionstats` show the info on a separate per-sstable compaction level, which is not good enough if you want to know if the whole...

So, will `nodetool compact`. Opening a separate GH.

> On Thu, Apr 1, 2021 at 10:48 AM Vladislav Zolotarov < ***@***.***> wrote: So, will nodetool compact. Opening a separate GH. > i think we possibly could make it...

> In general I'd like to see something more RESTy... > > At this point we should be consistent with what we have. > New repair status API `/storage_service/repair_status` with...

On a second thought "pausing"/"resuming" compactions implementation may be very not trivial since it may require locking the current sstables set for unknown time on the disk. So, we can...

> On Thu, Apr 1, 2021 at 12:28 PM Vladislav Zolotarov ***@***.***> wrote: On a second thought "pausing"/"resuming" compactions implementation may be very not trivial since it may require locking...

> We should skip to send update if the change is tiny. @asias why not to send the diff? This way you may not skip anything - if there was...

> IMO, it is better to generate less info in a smarter way than introducing a more powerful transfer channel. @asias We hit this issue (and tried to fix it)...

Oh, I was confused then. Let me re-review it.