What are your ARC statistics?

[Deadringers]

I'm being thick...but what do I put in the ""[*]Put your data type(s) here..."" section?

Can someone give me an example?

 

[Deadringers]

oh wait I really was being thick.....it literally just echo's this back to you so you can just put what ever you like.



Here are my stats - only 32GB of RAM installed right now, 128 GB to be installed today.

Not sure why it's not picking up any stats...thoughts?

• Put your data type(s) here...
• 10:45AM up 4 days, 13:08, 1 user, load averages: 0.11, 0.08, 0.09
• 680MiB / 28.5GiB (freenas-boot)
• 1.10TiB / 4.34TiB (vm-store-1)
• (MRU: , MFU: ) /
• Hit ratio -> (higher is better)
• Prefetch -> (higher is better)
• Hit MFU:MRU -> : (higher ratio is better)
• Hit MRU Ghost -> (lower is better)
• Hit MFU Ghost -> (lower is better)

 

[Deadringers]

Me being thick again, not the entire script was copied over - here are the new stats:

• Put your data type(s) here...
• 10:50AM up 4 days, 13:13, 1 user, load averages: 0.05, 0.09, 0.09
• 680MiB / 28.5GiB (freenas-boot)
• 1.09TiB / 4.34TiB (vm-store-1)
• 22.71GiB (MRU: 24.62GiB, MFU: 636.60MiB) / 32.00GiB
• Hit ratio -> 64.19% (higher is better)
• Prefetch -> 2.04% (higher is better)
• Hit MFU:MRU -> 21.25%:77.79% (higher ratio is better)
• Hit MRU Ghost -> 0.06% (lower is better)
• Hit MFU Ghost -> 1.20% (lower is better)

 

[Deadringers]

Okay after shutting down freenas, putting in 128GB of ram, and then booting it back up and copying over a load of VMs, here are the stats!

I expect it to get better over the next few days.

• Put your data type(s) here...
• 6:22PM up 31 mins, 1 user, load averages: 1.81, 2.19, 1.66
• 680MiB / 28.5GiB (freenas-boot)
• 1.12TiB / 4.34TiB (vm-store-1)
• 105.39GiB (MRU: 102.51GiB, MFU: 2.88GiB) / 128.00GiB
• Hit ratio -> 48.21% (higher is better)
• Prefetch -> 31.78% (higher is better)
• Hit MFU:MRU -> 67.68%:31.25% (higher ratio is better)
• Hit MRU Ghost -> 0.01% (lower is better)
• Hit MFU Ghost -> 0.00% (lower is better)

 

[Deadringers]

Just under a day of up-time and here's where I'm at.

• General VM store for about 80 VMs in ESXI
• 12:17PM up 18:26, 2 users, load averages: 0.06, 0.05, 0.06
• 680MiB / 28.5GiB (freenas-boot)
• 1.12TiB / 4.34TiB (vm-store-1)
• 88.47GiB (MRU: 79.96GiB, MFU: 21.72GiB) / 128.00GiB
• Hit ratio -> 89.18% (higher is better)
• Prefetch -> 49.90% (higher is better)
• Hit MFU:MRU -> 71.28%:22.42% (higher ratio is better)
• Hit MRU Ghost -> 0.00% (lower is better)
• Hit MFU Ghost -> 0.05% (lower is better)

 

[Deadringers]

at 3 days of uptime with same workload:

• Put your data type(s) here...
• 4:51PM up 2 days, 23 hrs, 2 users, load averages: 0.08, 0.07, 0.02
• 681MiB / 28.5GiB (freenas-boot)
• 1.13TiB / 4.34TiB (vm-store-1)
• 104.21GiB (MRU: 98.17GiB, MFU: 9.09GiB) / 128.00GiB
• Hit ratio -> 72.41% (higher is better)
• Prefetch -> 19.72% (higher is better)
• Hit MFU:MRU -> 43.75%:49.57% (higher ratio is better)
• Hit MRU Ghost -> 0.32% (lower is better)
• Hit MFU Ghost -> 1.21% (lower is better)

Interesting its getting "worse".

 

[Deadringers]

28 days uptime....

• Virtual Machine Storage
• 9:16AM up 28 days, 15:25, 1 user, load averages: 0.13, 0.06, 0.06
• 685MiB / 28.5GiB (freenas-boot)
• 745GiB / 4.34TiB (vm-store-1)
• 98.85GiB (MRU: 91.31GiB, MFU: 13.08GiB) / 128.00GiB
• Hit ratio -> 62.63% (higher is better)
• Prefetch -> 11.42% (higher is better)
• Hit MFU:MRU -> 26.36%:67.89% (higher ratio is better)
• Hit MRU Ghost -> 1.29% (lower is better)
• Hit MFU Ghost -> 6.04% (lower is better)

 

[Deadringers]

Same data set but with Dedup turned on....

• Virtual Machine Storage
• 9:14AM up 33 days, 15:23, 1 user, load averages: 0.53, 0.35, 0.34
• 690MiB / 28.5GiB (freenas-boot)
• 291GiB / 4.34TiB (vm-store-1)
• 101.65GiB (MRU: 67.81GiB, MFU: 34.98GiB) / 128.00GiB
• Hit ratio -> 70.18% (higher is better)
• Prefetch -> 11.23% (higher is better)
• Hit MFU:MRU -> 35.31%:55.52% (higher ratio is better)
• Hit MRU Ghost -> 0.95% (lower is better)
• Hit MFU Ghost -> 3.95% (lower is better)

 

[Recown]

I am realy concern about my result. Am i right, that result is so bad because i have pretty big pool and only 16GiB of memory? L2ARC is 80GiB SSD.


[*] Media data
[*] 5:15PM up 139 days, 23:52, 1 user, load averages: 0.05, 0.04, 0.05
[*]1023MiB / 74.5GiB (freenas-boot)
[*]18.5TiB / 29TiB (tank)
[*]50.70GiB (MRU: 4.40GiB, MFU: 46.30GiB) / 16.00GiB
[*]Hit ratio -> 28.99% (higher is better)
[*]Prefetch -> 17.65% (higher is better)
[*]Hit MFU:MRU -> 22.08%:39.90% (higher ratio is better)
[*]Hit MRU Ghost -> 79.64% (lower is better)
[*]Hit MFU Ghost -> 14.30% (lower is better)

 

[MrToddsFriends]

I am realy concern about my result.

No need to be concerned. FreeNAS versions prior to 9.10.2 showed a behavior that lead to spuriously high ARC hit ratios.

https://forums.freenas.org/index.php?threads/solved-differing-arc-usage-in-freenas-9-10-2.48831/

 

[Patrik Hansson]

Does this work on 11 ?
Only getting 1 row:

time read miss miss% dmis dm% pmis pm% mmis mm% arcsz c
16:17:26 0 0 0 0 0 0 0 0 0 10G 10G



 

[RoboKaren]

Same data set but with Dedup turned on....

You didn't get the memo that turning DEDUP ON is generally a bad idea?

 

[rogerh]

You didn't get the memo that turning DEDUP ON is generally a bad idea?

His system seems to meet the criteria for making it work, though.

 

[Scharbag]

My stats since running on a much less powerful system:

• Movies, Data, VMs shared over NFS to ESXi
  
• 11:34AM up 7 days, 23:58, 1 user, load averages: 0.26, 0.36, 0.45
  
• 25.9TiB / 32.5TiB (backuptank)
  
• 28.7TiB / 43.5TiB (bigtank)
  
• 144GiB / 444GiB (fasttank)
  
• 3.95GiB / 15.9GiB (freenas-boot)
  
• 11.07GiB (MRU: 8.91GiB, MFU: 2.19GiB) / 24.00GiB
  
• Hit ratio -> 28.90% (higher is better)
  
• Prefetch -> 43.12% (higher is better)
  
• Hit MFU:MRU -> 40.17%:23.92% (higher ratio is better)
  
• Hit MRU Ghost -> 0.87% (lower is better)
  
• Hit MFU Ghost -> 1.86% (lower is better)
  

Really would like to take the plunge and get a 64GB capable system...

OR grab a set and give the new version a chance and get rid of ESXI. But that is a TON of work to get my VMs transferred across...

Cheers,

 

[Hazimil]

• DVD/Blu-Ray Rips, User Data, MP3s, server currently in TEST mode, not live yet!
• 9:38AM up 20:21, 1 users, load averages: 0.01, 0.04, 0.06
• 513GiB / 58TiB (Tank)
• 993MiB / 111GiB (freenas-boot)
• 115.75GiB (MRU: 115.79GiB, MFU: 267.14MiB) / 128.00GiB
• Hit ratio -> 68.06% (higher is better)
• Prefetch -> 96.30% (higher is better)
• Hit MFU:MRU -> 53.13%:43.21% (higher ratio is better)
• Hit MRU Ghost -> 0.00% (lower is better)
• Hit MFU Ghost -> 2.61% (lower is better)

 

[Geoff Bricker]

Mine doesnt work and it saddens me dearly....

Show : What it looks like

root@ultron:/mnt/tank/HOME/test-scripts # ./ARCstats

• Put your data type(s) here...
• 3:12AM up 7 days, 18:58, 1 users, load averages: 0.29, 0.19, 0.11
• 2.88GiB / 14.9GiB (freenas-boot)
• 510GiB / 14.5TiB (tank)
• 11.0TiB / 29TiB (tanktwo)
  usage: awk [-F fs] [-v var=value] [-f progfile | 'prog'] [file ...]
  ./ARCstats: {print $6 $7}: not found
  usage: awk [-F fs] [-v var=value] [-f progfile | 'prog'] [file ...]
  ./ARCstats: {print $4}: not found
  usage: awk [-F fs] [-v var=value] [-f progfile | 'prog'] [file ...]
  ./ARCstats: {print $4}: not found
  usage: awk [-F fs] [-v var=value] [-f progfile | 'prog'] [file ...]
  ./ARCstats: {print $4}: not found
  usage: awk [-F fs] [-v var=value] [-f progfile | 'prog'] [file ...]
  ./ARCstats: {print $5}: not found
• 26.69 (MRU: , MFU: 8.19GiB) / 32.00
• Hit ratio -> Cache Hit Ratio: 70.39% 53.43m (higher i s better)
• Prefetch -> (higher is better)
• Hit MFU:MRU -> : (higher ratio is better)
• Hit MRU Ghost -> (lower is better)
• Hit MFU Ghost -> 2.05% (lower is better)

 

[Afrojoe]

• ESXi Datastore (5 VMs)
• 1:35PM up 19 days, 23:41, 1 users, load averages: 0.32, 0.33, 0.29
• 3.00GiB / 111GiB (freenas-boot)
• 1.10TiB / 18.1TiB (savage)
• 34.27GiB (MRU: 29.49GiB, MFU: 4.84GiB) / 40.00GiB
• Hit ratio -> 94.26% (higher is better)
• Prefetch -> 93.44% (higher is better)
• Hit MFU:MRU -> 56.43%:37.74% (higher ratio is better)
• Hit MRU Ghost -> 0.04% (lower is better)
• Hit MFU Ghost -> 0.13% (lower is better)

 

[clusty]

• Movies, Timemachine backups
• 10:04PM up 8 days, 51 mins, 0 users, load averages: 0.55, 0.47, 0.45
• 891GiB / 4.54TiB (data)
• 3.24GiB / 28.2GiB (freenas-boot)
• 8.30GiB (MRU: 5.47GiB, MFU: 2.83GiB) / 32.00GiB
• Hit ratio -> 66.23% (higher is better)
• Prefetch -> 27.36% (higher is better)
• Hit MFU:MRU -> 59.74%:33.95% (higher ratio is better)
• Hit MRU Ghost -> 0.78% (lower is better)
• Hit MFU Ghost -> 1.88% (lower is better)

 

[boynep]

• Media Storage(HomeStorage) with 4 X 3TB RaidZ1 Test is 4 X 120GB SSD raidZ1
  
• 9:48PM up 14 days, 4:10, 1 user, load averages: 0.51, 0.62, 0.56
  
• 5.98TiB / 10.9TiB (HomeStorage)
  
• 4.66GiB / 436GiB (Test)
  
• 4.45GiB / 7.06GiB (freenas-boot)
  
• 9.67GiB (MRU: 5.64GiB, MFU: 4.15GiB) / 16.00GiB
  
• Hit ratio -> 64.47% (higher is better)
  
• Prefetch -> 22.76% (higher is better)
  
• Hit MFU:MRU -> 89.36%:8.57% (higher ratio is better)
  
• Hit MRU Ghost -> 0.13% (lower is better)
  
• Hit MFU Ghost -> 0.55% (lower is better)
  

 

[BarefootWoodworker]

• NVR, ISO storage, and VM Storage. 19 VMs running off the VM iSCSI, about 15 of those get backed up nightly via BareOS + 2 offsite servers
• 12:52AM up 37 days, 3:11, 1 user, load averages: 0.58, 0.58, 0.57
• 2.97TiB / 3.62TiB (CCTV_iSCSI)
• 1.20TiB / 3.62TiB (Static_iSCSI)
• 6.45TiB / 14.5TiB (VM_iSCSI)
• 1.71GiB / 14.6GiB (freenas-boot)
• 107.20GiB (MRU: 82.32GiB, MFU: 24.88GiB) / 128.00GiB
• Hit ratio -> 83.46% (higher is better)
• Prefetch -> 34.52% (higher is better)
• Hit MFU:MRU -> 88.92%:10.30% (higher ratio is better)
• Hit MRU Ghost -> 0.31% (lower is better)
• Hit MFU Ghost -> 5.48% (lower is better)