1 Tests to evaluate the importance of various BlueSky features, so
2 different features are turned off for the benchmark runs.
4 20110305-west-fullsegments-4G
5 Test to S3-US-West, 4 GB disk cache, but all fetches from S3 are full
6 segments instead of byte ranges, and the complete segments are stored to
7 the disk cache. So there are multiple effects: first, for a random
8 workload there is more garbage data taking up cache space; second,
9 fetches from S3 take longer which increases read latencies.
14 BlueSkyRCStr *str = NULL;
16 size_t file_offset = 0, file_size = 0;
17 - gboolean range_request = TRUE;
18 + gboolean range_request = FALSE;
21 page_size = getpagesize();
23 Sat Mar 05 17:20:01 PST 2011
24 Executing run 1 of 25 ... done
25 Sat Mar 05 17:32:59 PST 2011
26 Executing run 2 of 25 ... done
27 Sat Mar 05 17:47:27 PST 2011
28 Executing run 3 of 25 ... done
29 Sat Mar 05 18:01:20 PST 2011
30 Executing run 4 of 25 ... done
31 Sat Mar 05 18:14:38 PST 2011
32 Executing run 5 of 25 ... done
33 Sat Mar 05 18:28:56 PST 2011
34 Executing run 6 of 25 ... done
35 Sat Mar 05 18:43:20 PST 2011
36 Executing run 7 of 25 ... done
37 Sat Mar 05 18:56:59 PST 2011
38 Executing run 8 of 25 ... done
39 Sat Mar 05 19:11:18 PST 2011
40 Executing run 9 of 25 ... done
41 Sat Mar 05 19:25:20 PST 2011
42 Executing run 10 of 25 ... done
43 Sat Mar 05 19:39:26 PST 2011
44 SfsManager shutting down ...
46 Cloudlog cache: 0 dirty, 0 writeback, 0 journal, 1734715 cloud
47 All segments have been flushed, journal < 12924 is clean
48 Ending cache size: 4191588 kB
51 Store[s3:mvrable-bluesky-west]: GETS: count=21781 sum=88260767105
52 Store[s3:mvrable-bluesky-west]: PUTS: count=13788 sum=55866496509
53 NFS RPC Messages In: count=2741556 sum=52779236892
54 NFS RPC Messages Out: count=2741556 sum=1931572500