Post Reply 
 
Thread Rating:
  • 0 Votes - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Monitoring performance / cache hit rate of SSD
11-25-2012, 04:52 PM
Post: #1
Monitoring performance / cache hit rate of SSD
I'm curious if there is a way to monitor the performance of an SSD in a 5D (in my case it's mSATA, but the question holds for any type of SSD on the 5D.)

Drobo hasn't described in detail how the SSD is used, so I'm not sure what the right questions are. Since it's working a block level, my guess is the most interesting data would be cached blocks and cache hit-rate. If Drobo is also using the SSD as a write-cache, the queue depth would also be interesting to monitor.

At the moment I'm getting poor performance out of the 5D, but that's because I'm using USB 2.0 on my Mac Pro until the new iMacs finally start shipping. I did some performance tests on my MacBook Pro via thunderbolt and that yielded good results.

Thanks,
Andrew
Find all posts by this user
Quote this message in a reply
11-25-2012, 05:27 PM
Post: #2
RE: Monitoring performance / cache hit rate of SSD
Sadly, Drobo tends not to expose even the most basic of info like SMART and SNMP diagnostics, so I doubt they'll ever let us at the nitty-gritty like SSD health and cache effectiveness.

They really just don't get the need for basic diagnostic data. It's not like any of this is a trade secret.

Drobo 5N | 250GB Samsung 850 EVO mSATA | 2 x 4TB Seagate, 3 x 4TB Hitachi | FS/EXT3 diskpack
Peak performance >100MBps read/write (based on FS disk pack, no jumbo frames, no SSD)
DroboPorts: Plex, Transmission, OpenSSH, NFS, nano, screen, bash
Find all posts by this user
Quote this message in a reply
11-25-2012, 05:30 PM (This post was last modified: 11-25-2012 05:34 PM by rdo.)
Post: #3
RE: Monitoring performance / cache hit rate of SSD
Drobo posted some details about how their new products use the SSD here: http://www.drobo.com/how-it-works/perfor...http://www.drobo.com/how-it-works/performance-profes

Also, I agree with diamondsw. Drobo just doesn't let us see that kind of information. (I had an amusing conversation with a tech who claimed that SMART status was a Drobo trade secret. I sent him a link to the published standards as evidence against his claim, and suggested that he provide supporting evidence before repeating such foolishness. Their culture is deep.)

You earn respect by how you live, not by what you demand.
-unknown
Find all posts by this user
Quote this message in a reply
02-20-2014, 05:17 AM
Post: #4
RE: Monitoring performance / cache hit rate of SSD
Hello all:

Sorry - I'm a bit late to the game here, but I'm currently evaluating the Drobo 5D to see if it's going to work with my particular application (audio sample playback, which runs lots of small random reads).

I'm trying to work out if the mSATA SSD is kicking in and improving the performance of my machine. As it's very difficult to get objective measurements from my audio applications, I'm simulating the process by running some shell scripts that first create 150GB of files, and then reading back small portions of some of these files repeatedly (for about an hour, then letting it rest to see if it needs to optimise, then trying again) - all whilst running iostat to monitor the performance.

Comparing the performance with the mSATA SSD in place, and without it in place, I can't see that the mSATA is improving the transfer speeds. Both with and without the SSD, I'm seeing around 300MB/s at present.

Can anyone give me a clue as to why these tests don't seem to be hinting at performance improvements with the SSD? From the published documentation, I would expect to see benefits with this transfer scenario.

Thanks.
Find all posts by this user
Quote this message in a reply
03-19-2014, 04:27 AM
Post: #5
RE: Monitoring performance / cache hit rate of SSD
No-one?
Find all posts by this user
Quote this message in a reply
03-19-2014, 10:27 AM
Post: #6
RE: Monitoring performance / cache hit rate of SSD
Typically the mSATA isn't going to affect your transfer speeds, it's going to keep them consistent so other I/O doesn't cause contention and stuttering behavior. At least that's been my experience with SSD caching.

Drobo 5N | 250GB Samsung 850 EVO mSATA | 2 x 4TB Seagate, 3 x 4TB Hitachi | FS/EXT3 diskpack
Peak performance >100MBps read/write (based on FS disk pack, no jumbo frames, no SSD)
DroboPorts: Plex, Transmission, OpenSSH, NFS, nano, screen, bash
Find all posts by this user
Quote this message in a reply
Post Reply 


Forum Jump: