Hi, been trying to run tilemaker (https://tilemaker.org/) on OpenBSD, and on 2 out of 2 runs (one was on 7.7 and the second one on 7.7-current) I am getting "pagedaemon: wait_pla deadlock detected!" with the machine becoming unresponsive (other than responding to ping) when cleaning up the process.
I am downloading https://download.geofabrik.de/europe-latest.osm.pbf and then running tilemaker on it, e.g. ./tilemaker --store temp-store europe-latest.osm.pbf europe.mbtiles the job seems to complete (as europe.mbtiles gets created successfully), but when cleaning up the process, I am getting "pagedaemon: wait_pla deadlock detected!" repeatedly printed on the console (and I can't login via ssh any more, it's just responding to pings). The machine has 32 GB of memory: real mem = 34292264960 (32703MB) avail mem = 33225953280 (31686MB) and CPU is Intel(R) Xeon(R) CPU E3-1240 v5 @ 3.50GHz, 3491.87 MHz, 06-5e-03, patch 000000f0 Note that I have also set "hw.smt=1" in /etc/sysctl.conf so I get 8 threads. I have changed the login class of the user to staff, so I can adjust the ulimit to: $ ulimit -a time(cpu-seconds) unlimited file(blocks) unlimited coredump(blocks) unlimited data(kbytes) 25165824 stack(kbytes) 4096 lockedmem(kbytes) 87381 memory(kbytes) 32425456 nofiles(descriptors) 512 processes 256 This time I had also run a "vmstat 60" (with added timestamps) in the background to get some idea of what is going on: 04:58:52 8 6626228M 1106M 4420 0 6 0 0 317 734 553 2143742 101757 82 15 2 04:59:52 1 5925657M 1336M 11985 0 1653 0 0 6832 5811 3082 1441219 72029 50 13 37 05:00:52 0 5925044M 1315M 27978 0 2073 6 0 23739 12219 6200 10 12368 0 8 92 05:01:52 0 5924755M 1141M 13456 0 917 18 0 10940 5863 3018 4 5994 0 4 96 05:02:52 0 5924394M 1138M 26358 0 2079 837 0 24901 13251 7294 4 14164 0 8 91 05:03:52 0 5923137M 1806M 25301 0 2057 1606 0 25775 14120 8338 4 14671 0 9 91 05:04:52 0 5924271M 74M 25809 0 1888 4456 0 23511 15442 9315 4 16534 0 8 91 05:05:56 0 5924432M 0M 687 0 68 564 0 565 597 496 15 884 0 0 100 05:07:03 0 5924433M 0M 6 0 0 21 0 21 12 85 19 163 0 0 100 05:07:52 0 5924433M 0M 4 0 0 2 0 2 1 78 18 143 0 0 100 Note that the timestamp on the generated file is 04:59:33 (and it seems to be intact). Any thoughts on this? Christof -- https://cmeerw.org sip:cmeerw at cmeerw.org mailto:cmeerw at cmeerw.org xmpp:cmeerw at cmeerw.org
