Roberto Villarreal
Roberto Villarreal
I was hoping that my previously observed behavior (that recovery required a host reboot) would still hold true. But a daemon restart (prior to swapping out binaries) resulted in the...
When I posted last night, those three hosts were approaching the 1k mark, and are now at 6k, 8k, and 11k. This is capture is with experiments still running, so...
It's still happening with the new binaries; all are steadily growing, with one of them crossing 1k already (and as I'm about to submit, reaching 2k). Proactive dump before it...
@tonistiigi We encountered this issue again under normal usage using the same release versions as my original report (but a different host). Because it was normal usage and was noticed...
Our CI has been on docker/moby 28.5.0 (which includes all the fixes/patches referenced here) for a week now, with no further occurrences of the issue and no reason to think...
Funny you should ask... there's a couple reasons why my docs PR is in draft mode. One of them was very related to this. When going through the original docs...
TL;DR; I think you can close this, though one of your earlier comments/ideas is still valid and should be explicitly decided. My original (potential) concern was `default = null` combined...
> varstring is the only one that can't be null for backwards compatibility. Agreed/understood. > But it might be too late to change that. I assume that was in the...
> but we can document that for this release In the actual docs, i.e. I should add a snippet in my documentation PR? Or something you'll take care of in...
Almost certainly related, but slightly different. That one mentions buildkit being intentionally configured with a bad value. In that scenario, a hang obviously isn't desirable, but it's understandable. In my...