New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
I will if I have no use for it.
Just spoke to support he says the only way I can get ipv6 working for a VM is to order a failover IP, which includes the Virtual MAC feature.
Without the Virtual MAC anytime you create a VM that creates a new MAC address the security of the switch will disconnect your server for an unauthorized MAC address..
He says it's pretty much impossible to use the ipv6 in a VM wihout this Virtual MAC.
Let me know when the time comes
Well you can give your VM the MAC address of the server and not have the server send any traffic on it's physical interface. That's what i did (using VMWare) and it's working fine.
I'm not sure if this will be of help to you since I have no experience with pfsense or any virtualization platform but...
Here's my story of a problem with online.net IPv6 network:
I had problems with the ipv6 in online.net as well. I managed to get it work on two servers with the exact server setup at first. Then I realized one of them lost ipv6 connectivity. Even though it got the ipv6 addresses it had no connection.
Then I tried to find a fix for it with what little I know. After a while I realized the only difference I had on these servers was that the one without ipv6 connectivity had my vpn server installed on it and I was distributing ipv6 addresses to my vpn clients from an online.net /64 subnet. The machines did get an IPv6 but refused connecting as well.
So, digging further I realized my machine didn't have a default route for IPv6. I was baffled, since they both had the exact ipv6 setup. After a bit more research I came to the conclusion that dnsmasq was causing the problem.
When there is another router advertising mechanism on the system my dedi stopped listening online.net network but listened to the internal network. Thus resulting in loss of connectivity.
I had to change
Entry in my /etc/network/interfaces to
In order to make it work again.
So, maybe it's something related to that in your case as well?
If not I'm sure someone like me will find this information useful. (:
So when I create the VM I just give it the MAC address of eth0? Kinda scared to try it lol..
I have so many issues right now with that server that I will probably remove everything and start over. Probably for the best since I've tested so many different things that it will never work.
You should make sure that the host server doesn't have an IP associated with eth0, so you'll have to do the configuration via iLO
In proxmox the IP is on vmbr0 but bridged to eth0
How the iLo Works ?? I turned it on but it shows unavailable on the mentioned IP. Then, I tried Rescue Mode. It doesn't get connected. How to use the KVM Over IP feature ?? I'm badly in need of partitioning it with LVM Storage.
@Mahfuz_SS_EHL Don't hesitate to open a ticket if you can't connect to your iLO
Were you on a VPn/Proxy connection whilst trying that? They do disable some functionality if you are so.
Rescue mode takes a while before you can connect.. give it say about 10 minutes or so before you can connect with putty.. and use port 22 for SSH..
Just wondering am I the only one getting this "port is off" error or is anyone else getting this?
I invite you to contact support by ticket and we can analyze.
Thanks for replying Vincent.. Yes been in contact with support since the day I purchased the server, from a failed hard drive to iLO being stuck now it's the network port getting cut off!! Not sure this time if it was just a short blip in the network cause I did not even do anything and just got disconnected.
Ok I will contact support again.. I've had my fair share of servers and I've never really need intervention from support as much as I have with this server lol..
Anyone has both SAS and SATA version? Can you compare the hard driver performance between SAS and SATA?
Don't have RAID 1 for the SATA.. but comparison for RAID 0
http://lowendtalk.com/discussion/comment/971850/#Comment_971850
@earl it seems the SAS is much faster than the SATA.
Oh, I see. I was hurrying :-P
No, I was on My Home Connection Direct.
Yes, SAS is 15K VS SATA which is 7200 RPM drive.. not sure about the cache size onboard the drives thought.
@Mahfuz_SS_EHL
I watched the boot process with iLO and realized that just booting the bios took a lot of time cause of all the tests..
Also If you reload proxmox from the console even thought you initially can SSH into the server you will get disconnected cause it will install proxmox on top of a debian install.. So it's best to give it sometime before you SSH to the server after a reload of proxmox..
Well, I am going to quote myself
After doing some benchmarks, I do find the results to be interesting:
Clearly the E3-1220 has a better Passmark score than the L3426 (6156 vs 3982)
But the hyperthreading does matter to Proxmox: the L3426 wins on both pveperf score and vzcpucheck score vs the E3-1220 (29791.76 vs 24744.32 and 744794 vs 618608, respectively)
The E3-1220 has faster RAM and scores higher regex/second via pveperf
The hard disk setup and amount of RAM is the same from the two offers
Based on Proxmox pveperf and vzcpucheck scores, it appears that the L3426 would be better for Proxmox virtualization.
The vzcpucheck is a simple script adding all your "bogomips" rows in /proc/cpuinfo and multiplying x 25 So, plain stupid IMHO.
And are you sure the pveperf test is better in that L3426? Because no, is plain perl, and it doesn't use multiple processes at all. And we are talking of processors with different architecture and really different clocks.
Also, as an aside note, the regex test isn't "per second" as it says, but "per 3 seconds" if you read the script.
So, imho, both programs are bad assessing virtualization performance.
OK, I have not examined the scripts. Just ran these so-called metrics + I know they are in line with actual experience running various server configs.
Yes, I am sure. I have actually run the scripts on all the hardware posted above as I currently have them all. In fact, I was doing to this to make some determinations on which to use for what purposes, etc.
And this would mean that the relative numbers still say the same given that that they are divided by a common denominator, correct?
This may very well be the case. YMMV and you need to compare to "real world" performance. For me, the real world performance has roughly mirrored the Proxmox numbers and, the numbers have been very informative in helping me determine what hardware to put into service for colo, etc.
Edit: BTW, it is like any other benchmark: Do you blindly trust the numbers? No. Can you combine it with other empirical data to discern some trends/patterns? Yes.
It is all an educated guess until you see what it does in the real world.
Weird results to be honest.
Of course, I just think their words are misleading.
Totally agree!!
Point me to any other benchmark you think would be more accurate so I can run it. I am seriously doing this to pick which to use for what. (Edit: And my primary need is virtualization)
Cheers
I have no idea if something like that exists. Now, I have two ideas:
First, more cores, equals to less contention ratio. Let's say, I have 8 cores to 1Ghz and maybe 4 cores at 3Ghz. Indeed, the second is faster at the end. But if I have 4 VMs running with 2 cores each one, in the second case they will be fighting for cpu (context switching) more than in the first case. So, the first case will have less strain on hjigh load environments.
Second, about the benchmark. In the case of VZ or LXC, since this isn't virtualization, any benchmark can help. Honestly I am a big hater of Unixbench, and a big fan of Geekbench (2.0, because 3.0 won't run on low ram environments). It's simple, super quick and tests several features of your processor, and just processor.
In the case of KVM, Vmware, maybe your best bet is to try benchmarks inside your VMs
And doing some googling, found this tool, of course I haven't tried it, but seems interesting:
http://www.vmware.com/products/vmmark/
Yup. The thing is all 4 of these servers (L3426, X3450, E3-1220v1, E3-1240v1) have 4 cores. All except the E3-1220v1 have 8 threads (E3-1220v1 has 4 threads; that's one of the big reasons for all the comparison).
I can run Geekbench, the question is: does it give any more accurate yardstick of virtualization performance (i.e., vs gaming, 3D rendering, compiling, etc on the baremetal)?? Does one benchmark translate directly to another?
The VMware tool looks interesting; and of course, requires installation of VMware
geekbench is very good ^
it's what i use to compare core performance usually.