Flash Pool perf data collection
2014-09-19 08:10 AM
OCUM 5.2R1 + 8.1.4P1 7-mode filers environment:
1. In the past we've been able to demonstrate the benefit of Flash Cache by looking at ext_cache_obj counter group, or going into wafl -> read_io_type and seeing how much load the FC card is taking on. Is there something that indicates how much IO is offloaded by SSD in flash pool? I'm wondering if hya_cache is it, but I also have a question about what the other hya_* counters are fo.
2. Is there a way to do this for write data?
3. How can we do this in cDOT? OCUM 5.2 for CDOT doesn't let us customize which counters to monitor like in 7-mode. OPM's interface is completely different and so far I haven't come up with a way to do that yet. I'll put some context around it. This isn't necessarily to convince storage admins(although we would appreciate some data around this). It's for when we're asked "do you really need to order SSDs instead of HDDs? They cost a lot more." I think everyone's heard that SSDs are faster than HDDs often enough, but it's good to be able to quantify it with our environment's data.