On 2024/10/21 01:12, William Thebrand wrote: > I pull the newest snapshot release every sunday, and run a full upgrade on my > system. Which > means the OS, packages, and even git based projects I use. Four times in a > row, the system has > crashed without warning, and is practically unusable. There is variance on > what I was doing > when the system decided to crash, so the source of the crash assumedly deals > with a background > process. > > I have tried to "tread lightly", and poke around to see if I can find the > error, but it just > crashes regardless, with little progress. As you would be aware of, with each > crash there is a > 30min filesystem check, which makes exploration rather time consuming.
Likely a locking change in uvm which has since been reverted. Either go back to the previous kernel or move forward. > This is the most severe bug I have come across in my experience with OpenBSD. > I can usually > work around them, but not this time. It's at a rather bad time, because I > need my system for > work. Generally I'd recommend having a backout plan if updating an important machine when it might be at a bad time if there were to be a problem with the new version.. > There are more photos than Gmail allows me to upload. So I might have to > creat a gallery in > google photos. That would be preferable to asking the mailing list server to send 34MB to each bugs@ subscriber anyway. In general retyping the panic message and at least the function names from the backtrace as text in the email will help in getting the right person to see the message in the first place. And providing URLs to screenshots will allow more detail without retyping without using as many resources at the list server or recipients.
