I've fixed this problem. It turns out that I accidentally deployed a bug fix branch onto a single server -- a branch, ironically, that contains an incomplete fix for a minor track bug, breaking track for other use cases.
Double-irony: All other clusters have automated deploy processes in place to prevent this sort of error, but not the cluster in question. Of course! -John Kalucki http://twitter.com/jkalucki Twitter, Inc. On Dec 1, 12:23 pm, John Kalucki <jkalu...@gmail.com> wrote: > I did a deploy at about that time, but only to tweek GC parameters. > Another developer has reported a similar issue. Digging... > > -John Kaluckihttp://twitter.com/jkalucki > Services, Twitter Inc. > > On Dec 1, 11:53 am, Mark McBride <mmcbr...@twitter.com> wrote: > > > > > > > Zac, can you let me know which userid you're using, the URL you're > > using to connect, and the time in UTC at which the error occurred? > > > ---Mark > > > On Tue, Dec 1, 2009 at 11:44 AM, Zac Witte <zacwi...@gmail.com> wrote: > > > Just before 8pm PST last night both my primary and secondary listening > > > servers stopped receiving updates and skipped notices and are only > > > getting keepalive messages. I tried restarting the process, but still > > > only keepalives. > > > > Was the stream turned off? Is it just me? > > > > Thx, > > > Zac > > > -- > > ---Mark > > >http://twitter.com/mccv