On 2017-11-21 18:21:16 -0500, Tom Lane wrote: > Justin Pryzby <pryzby@xxxxxxxxxxxxx> writes: > > As $subject: backends are stuck in startup for minutes at a time. I didn't > > strace this time, but I believe last time I saw one was waiting in a futex. > > Hm... A futex? Hm, that was stock postgres? > > I saved ~40 cores from backends from the most recent incident, which are all > > essentially identical: > > This one seems to be waiting for a shared LWLock (specifically, one of > the locks for the shared-buffer lookup hashtable), which could only block > if someone else is holding that lock exclusively, which implies that > there's at least one backtrace that's different from this. Hm. Wonder if that implies that there's a lot of cache pressure, or whether it's the compounded amount of How many different databases and roles are you having on that system? I guess iiiiiii is a sanitized name? Greetings, Andres Freund
