Traffic Graphic RAID 5 curious pattern

[gabrielpc1190]

Good day. I was backing up my esxi server to nas using a NFS share. When copying files I noted that the line in the traffic graphic go down each 30 seconds. I don't know why.
The destination volume is a RAID5 made of 3 320GB Seagate Sata Drives. Maybe is it normal from a RAID5 to do that?
I have attached a traffic graph screenshot to show you what I am talking about.
Anyone know something that can cause this?
Thank

My FreeNAS:
0.7.2 Sabanda (revision 8191)
Intel Pentium Dual Core 2.9 Ghz
DDR2 800Mhz 2GB RAM
RAID5 Hard Drives:
3 Seagate ST3320418AS-CC37-305245MB
RAID1 Hard Drives:
1 WDC-WD10EAVS-32D7B1-01.01A01
1 Hitachi-HDT721010SLA360-ST6OA3AA

RAID5 Info:
Geom name: RAID5
State: COMPLETE HOT
Status: Total=3, Online=3
Type: AUTOMATIC
Stripesize: 131072
MemUse: 483328 (msl 43)
Newest: -1
ID: 2314258937
Providers:
1. Name: raid5/RAID5
Mediasize: 640145686528 (596G)
Sectorsize: 512
Mode: r1w1e2
Consumers:
1. Name: ad10
Mediasize: 320072933376 (298G)
Sectorsize: 512
Mode: r2w2e3
DiskNo: 2
Error: No
2. Name: ad8
Mediasize: 320072933376 (298G)
Sectorsize: 512
Mode: r2w2e3
DiskNo: 1
Error: No
3. Name: ad6
Mediasize: 320072933376 (298G)
Sectorsize: 512
Mode: r2w2e3
DiskNo: 0
Error: No

 

[paleoN]

To quote myself:

If you have FreeNAS .7 you are in the wrong Forum. FreeNAS .7 has been rebranded to NAS4Free.

The forums you are looking for are located at http://forums.nas4free.org/. While you are there give them a hand.

 

[gabrielpc1190]

Thank you very much. I was ignoring this.

Good day. I was backing up my esxi server to nas using a NFS share. When copying files I noted that the line in the traffic graphic go down each 30 seconds. I don't know why.
The destination volume is a RAID5 made of 3 320GB Seagate Sata Drives. Maybe is it normal from a RAID5 to do that?
I have attached a traffic graph screenshot to show you what I am talking about.
Anyone know something that can cause this?
Thank

My FreeNAS:
0.7.2 Sabanda (revision 8191)
Intel Pentium Dual Core 2.9 Ghz
DDR2 800Mhz 2GB RAM
RAID5 Hard Drives:
3 Seagate ST3320418AS-CC37-305245MB
RAID1 Hard Drives:
1 WDC-WD10EAVS-32D7B1-01.01A01
1 Hitachi-HDT721010SLA360-ST6OA3AA

RAID5 Info:
Geom name: RAID5
State: COMPLETE HOT
Status: Total=3, Online=3
Type: AUTOMATIC
Stripesize: 131072
MemUse: 483328 (msl 43)
Newest: -1
ID: 2314258937
Providers:
1. Name: raid5/RAID5
Mediasize: 640145686528 (596G)
Sectorsize: 512
Mode: r1w1e2
Consumers:
1. Name: ad10
Mediasize: 320072933376 (298G)
Sectorsize: 512
Mode: r2w2e3
DiskNo: 2
Error: No
2. Name: ad8
Mediasize: 320072933376 (298G)
Sectorsize: 512
Mode: r2w2e3
DiskNo: 1
Error: No
3. Name: ad6
Mediasize: 320072933376 (298G)
Sectorsize: 512
Mode: r2w2e3
DiskNo: 0
Error: No

 

[survive]

Hi gabrielpc1190,

Odds are that's what they call "zfs breathing". PaeloN wrote a post on it here: http://forums.freenas.org/showthread.php?7473-ZFS-Breathing-Write-Stalls

On a side note......I dearly miss that realtime traffic graph from 0.7.X.

-Will

 

[paleoN]

This is the first thing that jumped out at me:

My FreeNAS:
0.7.2 Sabanda (revision 8191)

At which point I stopped reading. ;)

Odds are that's what they call "zfs breathing". PaeloN wrote a post on it here: http://forums.freenas.org/showthread.php?7473-ZFS-Breathing-Write-Stalls

Glancing at the graph I would say survive is spot on. Older versions of ZFS require a txg commit every 30 seconds. The new default as of ZFS v28 is every 5 seconds. Just changing that alone will likely smooth out your transfers quite a bit.