Dominic Evans

Fixed by https://github.com/Shopify/sarama/pull/2339 (thanks [Jacob-bzx](https://github.com/Jacob-bzx)) — will be included in the next release version

@KeenCN this is the first time I have seen this document, and I'd have to search to see if whoever is responsible for it ever raised any issues for these...

We're seeing a similar periodic issue on containerd v1.4.9 (Revision: e25210fe30a0a703442421b0f60afac609f950a3). After scaling up new nodes in an existing kube cluster (1.20.9), sometimes pods that are immediately scheduled onto that...

@AkihiroSuda / @fuweid is there any more we can collect for you here? any debug or metrics from the snapshotter that would be useful?

@fuweid did you get a chance to investigate any further on this? In the interim, are there a simple set of crd/crictl commands that we can use to delete the...

@fuweid in our case at least it is definitely not caused by a reboot, because it usually happens when a brand new machine is brought online as a replacement and...

Yeah I’ll need to grab some logs the next time it happens to see what the situation is with those two

@mvdan the proposal was to have an (optional) cmdline arg which would make shfmt attempt to wrap long lines to fit within 80 columns, line breaking at natural break points...

Cool, sounds reasonable. Closing.

@VipulZopSmart so EOF (clean socket shutdown) and broken pipe (unclean socket shutdown) errors are likely to be caused by your managed kafka service shutting down / restarting brokers/nodes or some...