Hi fio forum, Could you please address the following query: I was wondering about the correct way to perform a data integrity check while using a high stress scenario, Scenario 1: File system, Could the following command be used for data integrity check of IO on device with FS: fio --name=dedup --ioengine=libaio --rw=randwrite --bs=16k --numjobs=64 --thread --direct=1 --write_iolog=C --size=400m --directory=/mnt --do_verify=1 --verify=md5 I got the message that "Multiple writers may overwrite blocks that belong to other jobs. This can cause verification failures." In the fio man pages is written: filename=str Fio normally makes up a filename based on the job name, thread number, and file number. If you want to share files between threads in a job or several jobs, specify a filename for each of them to override the default Actually, fio creates 64 files and according to man pages every thread is writing to its file. Since -iodepth=1 (default) why should there be a problem with multiple writers even in the case of async ioengine? If this is not correct, can someone suggest a data integrity check command in multithreaded fio run for devices with FS? Scenario 2: RAW device, Could the flag --lockfile=readwrite in the following fio command solve the problem of "Multiple writers may overwrite blocks that belong to other jobs": fio --name=dedup --ioengine=posixaio --rw=randwrite --lockfile=readwrite --bs=32k --numjobs=64 --thread --size=100m --verify=md5 --do_verify=1 --direct=1 --filename=/dev/hdiskpower2 --verify_fatal=1 --invalidate=0 In the fio man pages is written: Lockfile=readwrite Read-write locking on the file. Many readers may access the file at the same time, but writes get exclusive access. Taking into account that that iodepth=1 and only one thread can write in a given time, one can suppose that it solves the problem of blocks overwrite since it looks similar to one threaded writer, isn't it? If this is not correct, can someone explain why and suggest a data integrity check command in multithreaded fio run for raw devices? Thank you Konstantin Ruvinsky -- To unsubscribe from this list: send the line "unsubscribe fio" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
