7 minutes to copy 4gb?
-
Do you believe that is a normal time to spend copying a 4 gb file to the same folder? I'm speaking of a linux server that has 3 SCSI 10K disks in RAID working together. The OS disks are also in RAID and are completely apart of the 3 data disks. What do you think?
That would be painfully slow in my Velociraptor SATA raid array on my pc, but I'm not sure about scsi these days.
"Creating your own blog is about as easy as creating your own urine, and you're about as likely to find someone else interested in it." -- Lore Sjöberg
-
Do you believe that is a normal time to spend copying a 4 gb file to the same folder? I'm speaking of a linux server that has 3 SCSI 10K disks in RAID working together. The OS disks are also in RAID and are completely apart of the 3 data disks. What do you think?
Sounds a little slow. Here is a linux software raid5:
datastore3 data_analysis # dd if=/dev/zero of=test.img bs=4M count=1000
1000+0 records in
1000+0 records out
4194304000 bytes (4.2 GB) copied, 17.5219 s, 239 MB/s
datastore3 data_analysis # echo 3 > /proc/sys/vm/drop_caches
datastore3 data_analysis # dd if=test.img of=test1.img bs=4M
1000+0 records in
1000+0 records out
4194304000 bytes (4.2 GB) copied, 50.1593 s, 83.6 MB/s
datastore3 data_analysis #John
-
Sounds a little slow. Here is a linux software raid5:
datastore3 data_analysis # dd if=/dev/zero of=test.img bs=4M count=1000
1000+0 records in
1000+0 records out
4194304000 bytes (4.2 GB) copied, 17.5219 s, 239 MB/s
datastore3 data_analysis # echo 3 > /proc/sys/vm/drop_caches
datastore3 data_analysis # dd if=test.img of=test1.img bs=4M
1000+0 records in
1000+0 records out
4194304000 bytes (4.2 GB) copied, 50.1593 s, 83.6 MB/s
datastore3 data_analysis #John
-
On one of my linux 7200 RPM SATA software raid 5s at work it took less then 51 seconds to copy a 4G file. Although I do admit this array is not just 3 drives. I believe it is 8 drives in raid 5.
John
-
Do you believe that is a normal time to spend copying a 4 gb file to the same folder? I'm speaking of a linux server that has 3 SCSI 10K disks in RAID working together. The OS disks are also in RAID and are completely apart of the 3 data disks. What do you think?
-
Do you believe that is a normal time to spend copying a 4 gb file to the same folder? I'm speaking of a linux server that has 3 SCSI 10K disks in RAID working together. The OS disks are also in RAID and are completely apart of the 3 data disks. What do you think?
-
That's just about 10mb/s which is kind of slow even when you are copying between the same drives, especially since you got RAID (configured for speed I assume).
If this is hardware raid 5 and the cache is disabled because of not having the battery unit this type of performance can happen.
John
-
Do you believe that is a normal time to spend copying a 4 gb file to the same folder? I'm speaking of a linux server that has 3 SCSI 10K disks in RAID working together. The OS disks are also in RAID and are completely apart of the 3 data disks. What do you think?
I can do about 40-50MB/sec between 2 SATA 150 drives. On the same drive it will be factors worse due to the obvious reasons.
xacc.ide
IronScheme - 1.0 RC 1 - out now!
((λ (x) `(,x ',x)) '(λ (x) `(,x ',x))) The Scheme Programming Language – Fourth Edition -
Do you believe that is a normal time to spend copying a 4 gb file to the same folder? I'm speaking of a linux server that has 3 SCSI 10K disks in RAID working together. The OS disks are also in RAID and are completely apart of the 3 data disks. What do you think?
I'll start this with the disclaimer that I know nothing about Linux file systems, but could it be that you are thrashing the drives bu moving the data from a low location on the disk to a high location on the disk (huge seek times), maybe compounded by doing this in small blocks? One way to eliminate this possibility is to copy the file to the other set of drives first, then back to the original set of drives, then divide the time by 2. Dave.
-
That would be painfully slow in my Velociraptor SATA raid array on my pc, but I'm not sure about scsi these days.
"Creating your own blog is about as easy as creating your own urine, and you're about as likely to find someone else interested in it." -- Lore Sjöberg
Showoff ;) would love to see a hdtach graph ;)
Personally, I love the idea that Raymond spends his nights posting bad regexs to mailing lists under the pseudonym of Jane Smith. He'd be like a super hero, only more nerdy and less useful. [Trevel]
| FoldWithUs! | sighist | µLaunch - program launcher for server core and hyper-v server -
Showoff ;) would love to see a hdtach graph ;)
Personally, I love the idea that Raymond spends his nights posting bad regexs to mailing lists under the pseudonym of Jane Smith. He'd be like a super hero, only more nerdy and less useful. [Trevel]
| FoldWithUs! | sighist | µLaunch - program launcher for server core and hyper-v serverpeterchen wrote:
would love to see a hdtach graph
I don't know where to put it so you can see it, but here are the stats from running it just now: Nvidia stripe 558.92G Random access: 7.0 ms CPU Utilization: 4% Average read: 173.8 MB/s Big red graph on the left: Burst speed - 374MB/s Strangely the stats were different a year ago when I first set this hardware up (worse). Maybe I did an update to something since then. It's xp pro 64 and I get errors when I save the test results in hdtach so I don't know if it's running properly or not, but it seems to be gathering the stats properly. I use this machine as nothing more than a host for my VMWare virtual development station and personal virtual workstation and I wanted to make sure that there was as little performance hit as possible by going virtual. I don't know if those numbers are any good or not but from my research a year ago I determined that performance should be good in this configuration. It might be quite out of date though by now.
"Creating your own blog is about as easy as creating your own urine, and you're about as likely to find someone else interested in it." -- Lore Sjöberg
-
peterchen wrote:
would love to see a hdtach graph
I don't know where to put it so you can see it, but here are the stats from running it just now: Nvidia stripe 558.92G Random access: 7.0 ms CPU Utilization: 4% Average read: 173.8 MB/s Big red graph on the left: Burst speed - 374MB/s Strangely the stats were different a year ago when I first set this hardware up (worse). Maybe I did an update to something since then. It's xp pro 64 and I get errors when I save the test results in hdtach so I don't know if it's running properly or not, but it seems to be gathering the stats properly. I use this machine as nothing more than a host for my VMWare virtual development station and personal virtual workstation and I wanted to make sure that there was as little performance hit as possible by going virtual. I don't know if those numbers are any good or not but from my research a year ago I determined that performance should be good in this configuration. It might be quite out of date though by now.
"Creating your own blog is about as easy as creating your own urine, and you're about as likely to find someone else interested in it." -- Lore Sjöberg
There's very litte information on how to configure a system to run virtual machines, the old adage "fast random access for OS, fast read / RAID for data" isn't that relevant anymore. So the current trend seems to be "make sure it's no bottleneck, whatever it costs" :) While the Velociraptor now gets beaten in data transfer rate, the random access time is still great - and seems to make a notable difference.
Personally, I love the idea that Raymond spends his nights posting bad regexs to mailing lists under the pseudonym of Jane Smith. He'd be like a super hero, only more nerdy and less useful. [Trevel]
| FoldWithUs! | sighist | µLaunch - program launcher for server core and hyper-v server