Unfortunately I don't have access to a pmem device but let's see how far we get: On 27 November 2017 at 12:39, Gavriliuk, Anton (HPS Ukraine) <anton.gavriliuk@xxxxxxx> wrote: > > result=$(fio --name=random-writers --ioengine=mmap --iodepth=32 --rw=randwrite --bs=64k --size=1024m --numjobs=8 --group_reporting=1 --eta=never --time_based --runtime=60 --filename=/dev/dax0.0 | grep WRITE) Please make your problem scenarios as simple as possible: 1. Just run fio normally so we can see the output it produces on both stdout and stderr normally. 2. Reduce the job back so it there are only the bare minimum options that reproduce the problem. 3. Try to avoid changing lots of things. Here you've switched ioengine introducing another place to look. Instead how about this: Was your /dev/dax0.0 device made using ndctl? Assuming yes: fio --name=dax-mmap --ioengine=mmap --rw=write --bs=64k --eta=never --time_based --runtime=60 --filename=/dev/dax0.0 --size=2g (apparently a size has to be specified when you try to use a character device - see https://nvdimm.wiki.kernel.org/ ) If you run just that by itself what do you see? Next up: fio --name=dax-dax --ioengine=dev-dax --rw=write --bs=64k --eta=never --time_based --runtime=60 --filename=/dev/dax0.0 --size=2g If you run just that by itself what do you see? Finally: Assuming a -o dax mounted filesystem on /pmem0/ : fio --name=libpmemblk --ioengine=pmemblk --rw=write --bs=64k --eta=never --time_based --runtime=60 --filename=/pmem0/fio-test,4096,1024 --thread=1 If you run just that by itself what do you see? (Perhaps the documentation for these ioengines and pmem devices needs to be improved?) -- Sitsofe | http://sucs.org/~sits/ -- 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
