Hi guys!
What is best practice for SMB?
What kind of rules do you put on outgoing traffic?
Do you use a http proxy?
Hi guys!
What is best practice for SMB?
What kind of rules do you put on outgoing traffic?
Do you use a http proxy?
With "NAS" are we talking about:
A file server is a NAS y'know.
You have a server in a chassi (SC846) just like the pic right? And with the two 2.5" drive bays option in the back next to the PSUs?
When you say sharing files with clients, over what medium? 10Gbe local LAN?
How are the backplane connected to the LSI controller? Is the LSI on the motherboard or PCIe card?
The reason I'm asking is to know what speeds we are talking about.
@krisleslie said in Purchase options for a small non profit 25 staff:
@pete-s said in Purchase options for a small non profit 25 staff:
uters and firewalls you'll find intel cpus with hardware accelerators doing the job. That way the cpu has the brute force to process a lot of packets but things that can be offloaded will be handed to the hardware accelerators.
True. If you want more power you'd have to get one of their two best routers. Which the top of the line infinity is decent. I've already considered moving to it
Yeah, that looks more like it.
Cavium that actually makes the CPU and SDKs have CPUs with up to 48 cores.
@jaredbusch said in Purchase options for a small non profit 25 staff:
Without QoS, the ERL can do nearly gigabit routing.
Almost all gigabit home routers can do that. It's because simple routing is hardware accelerated to a high degree.
But as soon as you need to do a little more with the traffic then it can't be hardware off-loaded, it has to be processed by the CPU and go the trough the entire ip stack in the CPU. And that's why a few options here and there can make such a dramatic difference on the throughput.
The ER-X for instance has a dual core 880 Mhz (0.8GHz) MIPS CPU.
You can see that it is very commonly used by asus, d-link, linksys, netgear etc in things like routers and APs. https://wikidevi.com/wiki/MediaTek_MT7621
In raw power that's like a decent smartphone from 2010.
If you look at more powerful routers and firewalls you'll find intel cpus with hardware accelerators doing the job. That way the cpu has the brute force to process a lot of packets but things that can be offloaded will be handed to the hardware accelerators.
@jaredbusch said in Purchase options for a small non profit 25 staff:
@pete-s said in Purchase options for a small non profit 25 staff:
@jaredbusch Wouldn't you get higher transfer rate if you disabled QoS? 66 Mbps is not impressive.
Maybe 10 years ago I bought the cheapest D-link home router I could find, i think it was $30 or something, for an emergency situation. It would push 95+ Mbps on a 100/10 fiber.
ANd how do your phone calls and such sound on that cheap D-Link? When 30 people are online at work?
Don't know. Back then we had ISDN for phones.
But the reason I mentioned it was just lack of knowledge from my part. I assumed that if a cheap home router could do 100Mbps of straight routing in 2008 then even a low cost enterprise router should breeze through 100Mbps of traffic shaping and what have you.
But I have never used an Ubiquiti Edgerouter so I thought they were more powerful. Looking closer I now see they have very modest hardware specs so then 66Mbps of QoS makes complete sense.
@jaredbusch Wouldn't you get higher transfer rate if you disabled QoS? 66 Mbps is not impressive.
Maybe 10 years ago I bought the cheapest D-link home router I could find, i think it was $30 or something, for an emergency situation. It would push 95+ Mbps on a 100/10 fiber.
@dbeato said in Do I have a power supply problem?:
Seeing the dates, is there a pattern? Like does it keep on going up to this year, 2018?
No, these were the only dates.
But I will update the bios/ipmi just in case it is a software problem as @momurda suggested.
Come to think of it, I'm not even sure the -12V in the power supply is even used anymore.
@jaredbusch said in Comparing 15k SAS and SSD:
@pete-s said in Comparing 15k SAS and SSD:
@obsolesce said in Comparing 15k SAS and SSD:
@travisdh1 said in Comparing 15k SAS and SSD:
@obsolesce Nice to confirm what most of us have been saying already.
Yeah its nice to see the numbers.
There is also the reliability rate to consider. SSD drives have 3 to 4 times lower annualized failure rate.
And you're more likely to use more 15K drives and do RAID 10 instead of RAID 1 with SSD.Regarding SAS/SATA SSD they will all disappear soon. NVME drives in the U.2 format have superior performance and a small price premium per GB compared to the same drive in SATA.
As far as I'm concerned SATA/SAS SSD should be bought for legacy use only.
That is totally not useful as the server manufacturers are not manufacturing servers to use those generally. Once that changes then I would recommend switching to those.
That might have been true in the past. But just looking at Dell Poweredge servers with one or two sockets, the following support hot-swap NVME drives : R6415, R7415, C4140, R440, R740xd, R640, R7425.
@obsolesce said in Comparing 15k SAS and SSD:
@travisdh1 said in Comparing 15k SAS and SSD:
@obsolesce Nice to confirm what most of us have been saying already.
Yeah its nice to see the numbers.
There is also the reliability rate to consider. SSD drives have 3 to 4 times lower annualized failure rate.
And you're more likely to use more 15K drives and do RAID 10 instead of RAID 1 with SSD.
Regarding SAS/SATA SSD they will all disappear soon. NVME drives in the U.2 format have superior performance and a small price premium per GB compared to the same drive in SATA.
As far as I'm concerned SATA/SAS SSD should be bought for legacy use only.
Nice! Brings me down memory lane.
But I have a little problem putting them up like that. Because the versions that has NT in their name was really the "Server" or "Workstation" edition. Desktops were running Windows 95/98/Me at the time.
I understand you are talking about the architecture but then maybe 2003/2008/2012/2016 server releases should also be there.
This is the story about NT and how it came about from one of the people involved:
https://web.archive.org/web/20100604082534/http://www.winsupersite.com/reviews/winserver2k3_gold1.asp
@momurda No, it never crashes. And it has dual power supplies.
I usually don't muck around much but I found these in the event log (IPMI webpage) on a Supermicro server.
What do you think? Is this something to worry about?
67 2017/11/15 19:37:46 -12V Voltage Lower Non-Critical - Going Low - Asserted
68 2017/11/15 19:37:46 -12V Voltage Lower Critical - Going Low - Asserted
69 2017/11/15 19:37:46 -12V Voltage Lower Non-Recoverable - Going Low - Asserted
70 2017/12/06 17:54:32 -12V Voltage Lower Non-Recoverable - Going Low - Deasserted
71 2017/12/06 17:54:32 -12V Voltage Lower Critical - Going Low - Deasserted
72 2017/12/06 17:54:32 -12V Voltage Lower Non-Critical - Going Low - Deasserted
73 2017/12/06 17:54:35 -12V Voltage Lower Non-Critical - Going Low - Asserted
74 2017/12/06 17:54:35 -12V Voltage Lower Critical - Going Low - Asserted
75 2017/12/06 17:54:35 -12V Voltage Lower Non-Recoverable - Going Low - Asserted
@momurda said in Web app authenticate against customer AD?:
@pete-s said in Web app authenticate against customer AD?:
But is it likely that an enterprise would expose ldap to the internet? Or is there something else inbetween?
You can do this, using ldaps and some certificates.
But likely you want to use SSO which is done over http/https. Many sites support SSO using SAML2.0 compliant implementations, like ADFS.
For example the MS CRM system has you setup an ADFS by default, you dont have to but is recommended, and i think required if you want remote users to use it without vpn.
This consists of the CRM, ADFS servers to provide access to people outside lan.
CRM homepage exposed to internet on 443, ADFS server on 443 exposed as well.Someone on outside network, they sign into CRM homepage with AD creds. The login request gets sent to public IP of your ADFS server over https, which then connects to you AD server on the LAN, it does its checks and responds with yay or nay to CRM.
Thanks, I'll look more into this.
But is it likely that an enterprise would expose ldap to the internet? Or is there something else inbetween?
@momurda said in Getting A Dell T320 To Work As Dual Monitor Workstation:
PCIe bus power is 75W without need for 6 or 8 pin power from GPU.
Yes, and almost all entry level cards are WAY below that.
@nashbrydges said in Getting A Dell T320 To Work As Dual Monitor Workstation:
No, trying to get dual mons working on Win 10 Pro. I could get a new card but then I have to make sure it's going to be under 65W (I think that's the max) then try to get it to work. He's got the Quadro already so trying to work with what he has. Like I said, I'm close to just letting him know he needs to suck it up and buy a workstation but thought this might be worth a try.
As usual Dell, HP and similar have strange limitations in their machines.
https://www.dell.com/community/PowerEdge-Hardware-General/T320-video-card-upgrade/td-p/4404427
PS. A lot of users find the tower servers too loud for desktop work.
I have an enterprise customer who uses web applications on the net where the users log in with their regular login / password from the customers AD.
How is this done?
XCP-ng, the community version of Xenserver, has released version 7.5.
https://xcp-ng.org/2018/08/10/xcp-ng-7-5/
@emad-r said in VM host: dual CPU vs single CPU - same CPU performance rating:
@pete-s
Also what is the defacto standard in bench-marking single CPU rating ?
How do you test ?
I use the Passmark CPU test because they have a benchmark you can run on your machine and then automatically submit the test results. Then you can see the average results on the website. https://www.cpubenchmark.net/
A CPU like Xeon E5-2630 V4 has the result submitted from 59 machines and then averaged.
The CPU test consists of the following
This is just a general benchmark for the CPU.