Thanks Sitsofe On 11/13/18, Sitsofe Wheeler <sitsofe@xxxxxxxxx> wrote: > Hi, > > (You just restated your original question verbatim which unfortunately > doesn't move the problem forward. I was hoping you were going to point > out the exact specific part of the offset modifier concept you were > struggling with by using a worked example to show what you did > understand and which part you didn't so we could expand on just that. > I'll take a stab at explaining things this time but next time try and > be specific and briefly demonstrate what bits you did understand if > you're asked for clarification :-) > > I'm just going to quote some text because "offset" in a different > context refers to another option: > > From > https://fio.readthedocs.io/en/latest/fio_doc.html#cmdoption-arg-readwrite : > "It is possible to specify the number of I/Os to do before getting a > new offset by appending :<nr> to the end of the string given. For a > random read, it would look like rw=randread:8 for passing in an offset > modifier with a value of 8. If the suffix is used with a sequential > I/O pattern, then the <nr> value specified will be added to the > generated offset for each I/O turning sequential I/O into sequential > I/O with holes. For instance, using rw=write:4k will skip 4k for every > write. Also see the rw_sequencer option." > > So in the above the "offset modifier/suffix" has two different meanings: > - With random I/O, the offset suffix is saying "do <nr> I/Os > sequentially before randomly* picking another starting point and then > doing another <nr> I/Os sequentially" > - With sequential I/O the offset suffix refers to the "distance to > skip between I/Os" > > Seen another way, depending on context the offset modifier is > controlling the "the offset (distance) chosen from the first allowable > position for I/O before doing sequential I/O" or the "offset > (distance) from the end of the previous I/O". > > (Note: I'm skipping over the effect of rw_sequencer in the above to > keep this discussion shorter) > > As to your second question, for many jobs so long as I use the same > job options verification should work with an offset modifier. For > example: > > fio --name=gappy_verify --size=512k --filename=/tmp/fio.tmp > --rw=write:4k --verify=crc32c > > wrote every other 4k and then verified it. If you're splitting up your > verifies into separate jobs for some reason, you may find the > verify_only and do_verify options useful (see > https://fio.readthedocs.io/en/latest/fio_doc.html#verification ). > > Does this help at all? > > * This is a simplification but it gives you the idea. > > On Tue, 13 Nov 2018 at 01:20, Mohanraj B <bmohanraj91@xxxxxxxxx> wrote: >> >> Hi, >> >> Thanks for the response. >> >> I have gone through and worked out the example. >> Below are my queries: >> what is offset? >> How to verify the results if used the :<nr> at the end of the string ? >> >> Thanks and Regards, >> Mohan >> >> On 11/7/18, Sitsofe Wheeler <sitsofe@xxxxxxxxx> wrote: >> > Hi, >> > >> > On Wed, 7 Nov 2018 at 11:31, Mohanraj B <bmohanraj91@xxxxxxxxx> wrote: >> >> >> >> Hello all, >> >> >> >> >> >> https://fio.readthedocs.io/en/latest/fio_man.html#i-o-type >> >> >> >> what is offset modifier, some please explain. >> >> It is possible to specify the number of I/Os to do before getting a >> >> new offset by appending :<nr> to the end of the string given. >> > >> > I guess you mean mean the rw <nr> parameter in the final paragraph in >> > https://fio.readthedocs.io/en/latest/fio_man.html#cmdoption-arg-readwrite >> > . Can you say what part of that paragraph and the rw_sequencer section >> > you are struggling to understand and did you work through the >> > randread:8 and rw=write:4k examples? >> > >> > -- >> > Sitsofe | http://sucs.org/~sits/ >> > > > -- > Sitsofe | http://sucs.org/~sits/ >
