100K ops/sec on the filer
We were under a bit of pressure to finish some analysis on a very tight deadline last week, so I opened up the rate limits between the NFS and Cluster networks and told the boys to “give ‘er”. I was blown away to see this:
A close up:
Ooooooh so close to crossing the 100,000 NFS ops/sec mark! As you can guess, the nodes were doing a lot of very small reads/writes, but we were still pushing over 300MB/sec. I really didn’t think you could push a FAS960 that hard, but there you have it. Other volumes definitely felt a little sluggish during the onslaught, but nobody actually noticed enough to complain about it.
On a related note, I uploaded two other shots of the filer to Flickr, and tagged them as “netapp“. Turns out I’m the first person to tag a photo with netapp, but I can see that more photos have since shown up on the public netapp photostream. Cool!
7 Comments
Jump to comment form | comments rss [?] | trackback uri [?]