== sanity test 255a: check 'lfs ladvise -a willread' ===== 11:40:56 (1781624456) 100+0 records in 100+0 records out 104857600 bytes (105 MB, 100 MiB) copied, 7.4437 s, 14.1 MB/s ladvise: range [2, 1] is invalid Provide servers with advice about access patterns for a file. usage: ladvise [--advice|-a ADVICE] [--start|-s START[kMGT]] [--background|-b] [--unset|-u] {--end|-e END[kMGT]|--length|-l LENGTH[kMGT]} {[--mode|-m [READ,WRITE]} ... ladvise: range [2, 2] is invalid Provide servers with advice about access patterns for a file. usage: ladvise [--advice|-a ADVICE] [--start|-s START[kMGT]] [--background|-b] [--unset|-u] {--end|-e END[kMGT]|--length|-l LENGTH[kMGT]} {[--mode|-m [READ,WRITE]} ... ladvise: conflicting arguments of -l and -e Provide servers with advice about access patterns for a file. usage: ladvise [--advice|-a ADVICE] [--start|-s START[kMGT]] [--background|-b] [--unset|-u] {--end|-e END[kMGT]|--length|-l LENGTH[kMGT]} {[--mode|-m [READ,WRITE]} ... Synchronous ladvise should wait fail_val=8 fail_loc=0x237 Asynchronous ladvise shouldn't wait llite.lustre-ffff9146cbd6a000.hybrid_io=0 Iter 1/10: reading without willread hint Iter 1/10: uncached speed: 0.572629 Iter 1/10: OSS cache speed: 0.619603 Iter 1/10: ladvise speed: 0.646841 Iter 2/10: reading without willread hint Iter 2/10: uncached speed: 0.641334 Iter 2/10: OSS cache speed: 0.646825 Iter 2/10: ladvise speed: 0.391156 Iter 3/10: reading without willread hint Iter 3/10: uncached speed: 0.301948 Iter 3/10: OSS cache speed: 0.364414 Iter 3/10: ladvise speed: 0.38848 Iter 4/10: reading without willread hint Iter 4/10: uncached speed: 0.366451 Iter 4/10: OSS cache speed: 0.386362 Iter 4/10: ladvise speed: 0.380384 Iter 5/10: reading without willread hint Iter 5/10: uncached speed: 0.42549 Iter 5/10: OSS cache speed: 0.456558 Iter 5/10: ladvise speed: 0.439565 Iter 6/10: reading without willread hint Iter 6/10: uncached speed: 0.458006 Iter 6/10: OSS cache speed: 0.476083 Iter 6/10: ladvise speed: 0.489402 Iter 7/10: reading without willread hint Iter 7/10: uncached speed: 0.465425 Iter 7/10: OSS cache speed: 0.462387 Iter 7/10: ladvise speed: 0.554395 Iter 8/10: reading without willread hint Iter 8/10: uncached speed: 0.527099 Iter 8/10: OSS cache speed: 0.531631 Iter 8/10: ladvise speed: 0.532008 Iter 9/10: reading without willread hint Iter 9/10: uncached speed: 0.499761 Iter 9/10: OSS cache speed: 0.543124 Iter 9/10: ladvise speed: 0.55009 Iter 10/10: reading without willread hint Iter 10/10: uncached speed: 0.578413 Iter 10/10: OSS cache speed: 0.579551 Iter 10/10: ladvise speed: 0.585219 Average uncached read: .48 Average speedup with OSS cached read: .50 = +4% Average speedup with ladvise willread: .49 = +2% Speedup with OSS cached read less than 20%, got 4%. Skipping ladvise willread check. fail_loc=0