Hi Sudheesh, I have been looking at this and the root cause is that the file system, being queries, was not available. This was easy to fix but the reaction of Drill could perhaps have been better (maybe it was stuck waiting for the filesystem) or I could shorten the timeout.
Anyways; I think this classifies as a "stupid user" case but I can mention it in Jira if you still think it's relevant. Thank you, -Stefan On Mon, Aug 3, 2015 at 4:48 PM, Sudheesh Katkam <skat...@maprtech.com> wrote: > Hi Stefan, > > Can you create a JIRA for this? Please attach to the JIRA: > (1) thread dumps of the three Drillbits (you can get this using jstack), > and > (2) json query profiles of the PENDING queries (you can get this from the > "Full JSON Profile" at the bottom of the profile page). > > Thank you, > Sudheesh > > > On Aug 3, 2015, at 9:00 AM, Stefán Baxter <ste...@activitystream.com> > wrote: > > > > Hi, > > > > I have a small cluster of 3 drillbits running. It's been working just > fine > > until it stopped working altogether. I notice a few "pending" queries and > > when I try to cancel them, via the admin, they either report that they > > don't know where they are running or the cancelling process freezes. > > > > What is a) the easiest way to delete all pending tasks? and b) make sure > > that this does not happen? > > > > All the best, > > -Stefan > >
