https://bugzilla.wikimedia.org/show_bug.cgi?id=68620
--- Comment #2 from Nemo <[email protected]> --- (In reply to Bawolff (Brian Wolff) from comment #1) > Wonder if its something specific to the labs-vagrant setup (e.g. Maybe a > network file system causes really slow writes, so writing to an sqlite db > takes a long time)? There was some IO wait but the biggest problem seemed to be insufficient RAM (!). That's an instance with only 2 GB and half is consumed by puppet, as said above. 21.30 < ori> [mw1053:~] $ sudo hdparm -Tt /dev/sda1 21.30 < ori> /dev/sda1: 21.31 < ori> Timing cached reads: 9628 MB in 1.99 seconds = 4830.60 MB/sec 21.31 < ori> Timing buffered disk reads: 318 MB in 3.02 seconds = 105.39 MB/sec 21.31 < ori> [mw1053:~] $ sudo hdparm -Tt /dev/sda1 21.31 < ori> /dev/sda1: 21.31 < ori> Timing cached reads: 10612 MB in 1.99 seconds = 5320.34 MB/sec 21.31 < ori> Timing buffered disk reads: 304 MB in 3.01 seconds = 100.89 MB/sec 21.31 < ori> mw1053:~# dd if=/dev/zero of=/tmp/output bs=8k count=10k; rm -f /tmp/output 21.31 < ori> 10240+0 records in 21.31 < ori> 10240+0 records out 21.31 < ori> 83886080 bytes (84 MB) copied, 0.144263 s, 581 MB/s 21.42 < Nemo_bis> convert and texvc hammering disk, according to htop 21.45 < Nemo_bis> 83886080 bytes (84 MB) copied, 1.06678 s, 78.6 MB/s / but on an idle instance 83886080 bytes (84 MB) copied, 0.0874509 s, 959 MB/s 21.49 < Nemo_bis> Though, when slightly bigger: $ dd if=/dev/zero of=/tmp/output bs=40k count=50k; rm -f /tmp/output / 2097152000 bytes (2.1 GB) copied, 36.9018 s, 56.8 MB/s -- You are receiving this mail because: You are the assignee for the bug. You are on the CC list for the bug. _______________________________________________ Wikibugs-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/wikibugs-l
