Open Metroidinator opened 3 years ago
If you can zip up the whole test directory that would help, assuming no proprietary info is in it (i.e you're not testing a prototype or something). The details dir has the raw CSV outputs used to generate the exceedance data, so it'd be necessary to see those to verify FIO actually wrote the needed data (FIO output format and options often change).
Attached is the output directory. details_15360GB_72cores_3000MHz_nvme0n1_fstlnsr650-8.fst.local_2021-03-09_14-03-52.zip
I've isolated the issue to this if block: https://github.com/earlephilhower/ezfio/blob/79909e6cce3f8dd6d1b0929156014d9c6d396d93/ezfio.py#L627-L630
Bypassing this conditional allows the *.exc.read/write.csv files to be generated (with data!). I don't know if ['clat_ns']['N'] still exists. I could not find it in the latest FIO documentation or through any of my debugging.
Hello, After running ezfio.py there was no exceedance data to look at or graph. Other data and values are present. No obvious errors or warnings in stdout. ezfio_exceedance_exceedance30.csv was generated without any data in it. After browsing through the code, it seems some *.exc.read/write.csv files should have been created but are not present.
Command line:
./ezfio.py -d /dev/nvme0n1 --yes
Environment:
CentOS Linux release 8.2.2004 (Core)
Linux fstlnsr650-8.fst.local 4.18.0-193.19.1.el8_2.x86_64 #1 SMP Mon Sep 14 14:37:00 UTC 2020 x86_64 x86_64
Python 3.6.8
fio-3.7
Please let me know if I can provide more information. Best Regards, Calvin