XenServer 7 has launched!
-
@BRRABill said in XenServer 7 has launched!:
@travisdh1 said in XenServer 7 has launched!:
@BRRABill 1Gb networks are much nicer, and not any more expensive anymore. For future consideration.
Yeah, it's a long story.
The servers are all connected via 1GB, though.
Assuming we both make it to Mangocon, maybe we can trade war stories
-
@travisdh1 said
Assuming we both make it to Mangocon, maybe we can trade war stories
Sounds good!
I mean, 100Mbps is fine for 99% of what we do. It's just these VM transfers are a bear.
But, the new me with my fancy ML knowledge now just sets up a Linux instance and does the import/export on the 1GB segment. Viva la open source!
Of course THAT was also slow, which has me wondering... But as I said, nothing firm tested yet so we shall see.
-
@Dashrender said in XenServer 7 has launched!:
@olivier said in XenServer 7 has launched!:
There is no magic: you need to have enough write speed on the target and read speed on the source.
What is the sustained write speed on a SATA drive? I'm guessing it must be a lot lower than 120 MB/s??
Way, way slower.
-
@travisdh1 said in XenServer 7 has launched!:
@BRRABill said in XenServer 7 has launched!:
@olivier said
There is no magic: you need to have enough write speed on the target and read speed on the source.
Yeah, I'll set up some tests once I am done getting everything set back up.
Right now it's clipping along at 5MB/s so it's taking forever.....
Something is wrong with that picture. Even older SATA drives can hit 100MB/s. 100MB network connection could be topping out at that.
Can hit, but not sustain. Generally can't hit for very long at all.
-
@BRRABill said in XenServer 7 has launched!:
For me, this is SSD going to SSD.
Granted, this current import is from my 100mbps desktop connection, but still, 40mbps is slow.
Don't mix speeds or write speed casually. Your desktop connection must be 100Mb/s (seriously, though, how did this happen, am I confused and this is still 2004?) and the drive speed is written in 40MB/s which is 320Mb/s. Right?
-
@travisdh1 said in XenServer 7 has launched!:
@BRRABill said in XenServer 7 has launched!:
For me, this is SSD going to SSD.
Granted, this current import is from my 100mbps desktop connection, but still, 40mbps is slow.
That's within reason for a 100Mb network connection. The best I've seen out of a 100Mb network is 9MB/sec, that's on a network with zero other traffic.
Yeah, that even would be a rare speed. The rule of thumb is that on an empty connection you'd unlikely break 8MB/s.
-
@scottalanmiller said in XenServer 7 has launched!:
@BRRABill said in XenServer 7 has launched!:
For me, this is SSD going to SSD.
Granted, this current import is from my 100mbps desktop connection, but still, 40mbps is slow.
Don't mix speeds or write speed casually. Your desktop connection must be 100Mb/s (seriously, though, how did this happen, am I confused and this is still 2004?) and the drive speed is written in 40MB/s which is 320Mb/s. Right?
Sigh - I installed new switches here in 2008 that were all 100 Mb - the price for 1 Gb switches was nearly double.
We are looking at replacing the switches, though like our old Wireless G system, they aren't failing, they are just on the slow side. And unlike the wireless - which is EOL, no support - the HP Switches have a lifetime guarantee.
So I find myself in the same quandry I asked about a year ago with my wireless - do I replace the switches just because they are old and slow, but not affecting production?
Sure in testing or me personally, I'm affected when I try to copy a large file to install Dragon Naturally Speaking, on the rare cases I need to do so, but is that enough to warrant switch replacement?
-
@Dashrender said in XenServer 7 has launched!:
Sigh - I installed new switches here in 2008 that were all 100 Mb - the price for 1 Gb switches was nearly double.
That's half a decade after I replaced the switches at HOME with GigE!
You know you could have put in 100Mb/s Hubs for free that would have been "infinitely" cheaper than the 100Mb/s ones.
Price ratio isn't important, ROI is. -
@Dashrender said in XenServer 7 has launched!:
So I find myself in the same quandry I asked about a year ago with my wireless - do I replace the switches just because they are old and slow, but not affecting production?
Sure in testing or me personally, I'm affected when I try to copy a large file to install Dragon Naturally Speaking, on the rare cases I need to do so, but is that enough to warrant switch replacement?
Are they really not affecting production? That's surprising. For a lot of stuff I can see that they would not, but it sure seems like from time to time that people would be experiencing latency from that. I sure would at home.
-
@scottalanmiller said in XenServer 7 has launched!:
Are they really not affecting production? That's surprising. For a lot of stuff I can see that they would not, but it sure seems like from time to time that people would be experiencing latency from that. I sure would at home.
Until recently they were all sharing a 10/10 internet pipe (now a 100/20 pipe).
99% of work is down through the internet, even at the faster speeds, no one is getting the whole 100 to themselves around here, so I wouldn't expect it to make much difference.
-
@Dashrender said in XenServer 7 has launched!:
@scottalanmiller said in XenServer 7 has launched!:
Are they really not affecting production? That's surprising. For a lot of stuff I can see that they would not, but it sure seems like from time to time that people would be experiencing latency from that. I sure would at home.
Until recently they were all sharing a 10/10 internet pipe (now a 100/20 pipe).
99% of work is down through the internet, even at the faster speeds, no one is getting the whole 100 to themselves around here, so I wouldn't expect it to make much difference.
@Dashrender said in XenServer 7 has launched!:
@scottalanmiller said in XenServer 7 has launched!:
Are they really not affecting production? That's surprising. For a lot of stuff I can see that they would not, but it sure seems like from time to time that people would be experiencing latency from that. I sure would at home.
Until recently they were all sharing a 10/10 internet pipe (now a 100/20 pipe).
99% of work is down through the internet, even at the faster speeds, no one is getting the whole 100 to themselves around here, so I wouldn't expect it to make much difference.
There is no file serving or anything internally?
-
@scottalanmiller said
Are they really not affecting production? That's surprising. For a lot of stuff I can see that they would not, but it sure seems like from time to time that people would be experiencing latency from that. I sure would at home.
Not at my office, no.
-
@scottalanmiller said
Don't mix speeds or write speed casually. Your desktop connection must be 100Mb/s (seriously, though, how did this happen, am I confused and this is still 2004?) and the drive speed is written in 40MB/s which is 320Mb/s. Right?
It was 40 mbps not MB/s
Assuming those are different things?
Megabits (m) versus Megabytes (MB)
Or do I have the case wrong?
-
Ah, I see I do. Thank you Google.
It was 40 Mbps. Or around 5 MBps.
-
So before I get into testing, let me explain what I have here. Perhaps I am setting something up wrong.
I have a pretty new DELL T320 server that is running XS7.
On this server, I have two test machines running Mint.
I am using iPerf between the two VMs to test bandwidth. I set up a test for 120 seconds, and got about an average of 110MBps. The performance iPerf reported matched the perrformance counter in XC and XO.
Is iPerf a resonable test of bandwidth speed, or no?
-
@BRRABill Yes. iPerf will give you the max network speed between the two testing points. Which type of network adapter is assigned to them?
-
@BRRABill said in XenServer 7 has launched!:
Ah, I see I do. Thank you Google.
It was 40 Mbps. Or around 5 MBps.
There you go
-
I then copied a 7GB file between the two VMs. That gives me a speed of around 30MBps. Nowehere near the iPerf numbers, but decently fast. (The file copied in 4 minutes.
I loaded XO on the same VMs, and tried an export of another VM. I am getting speeds of anywhere from 2MBps to 7MBps.
Clearly much slower.
-
@BRRABill Right. iPerf is testing nothing but the network. Think of it as the best you can get.
-
@travisdh1 said
@BRRABill Right. iPerf is testing nothing but the network. Think of it as the best you can get.
I am trying to get iPerf installed on my XS7 host. Being a Linux noob I am having some issues but working through them.
But I am assuming it is going to show the same results ... near GB speeds.
