What Are You Doing Right Now
-
@Dashrender said in What Are You Doing Right Now:
@WrCombs said in What Are You Doing Right Now:
@WrCombs said in What Are You Doing Right Now:
Creating a Zip Folder for my ISO of Server 2019 to put onto a flash drive so I can Use it at my house after I install Fedora 31 and KVM to practice Windows AD for hands on practice.
apparently even after creating a zip file it's too large to transfer..
To transfer to where? You'll need Fat32 up to a 4 GB file, you'll need NTFS for anything larger as a single file.
-
@WrCombs said in What Are You Doing Right Now:
@Dashrender said in What Are You Doing Right Now:
@WrCombs said in What Are You Doing Right Now:
@WrCombs said in What Are You Doing Right Now:
Creating a Zip Folder for my ISO of Server 2019 to put onto a flash drive so I can Use it at my house after I install Fedora 31 and KVM to practice Windows AD for hands on practice.
apparently even after creating a zip file it's too large to transfer..
To transfer to where? You'll need Fat32 up to a 4 GB file, you'll need NTFS for anything larger as a single file.
so you have to copy anything off of it you want to save - and then reformat it as NTFS
-
@Dashrender said in What Are You Doing Right Now:
@WrCombs said in What Are You Doing Right Now:
@Dashrender said in What Are You Doing Right Now:
@WrCombs said in What Are You Doing Right Now:
@WrCombs said in What Are You Doing Right Now:
Creating a Zip Folder for my ISO of Server 2019 to put onto a flash drive so I can Use it at my house after I install Fedora 31 and KVM to practice Windows AD for hands on practice.
apparently even after creating a zip file it's too large to transfer..
To transfer to where? You'll need Fat32 up to a 4 GB file, you'll need NTFS for anything larger as a single file.
so you have to copy anything off of it you want to save - and then reformat it as NTFS
Gotcha..
-
@Dashrender said in What Are You Doing Right Now:
@siringo said in What Are You Doing Right Now:
@scottalanmiller said in What Are You Doing Right Now:
@siringo said in What Are You Doing Right Now:
having to put in reservations for .255 & .0 ip addresses as windows dhcp issues them when you have a /23 scope. annoying.
Why would you be reserving them? In a /23 scope they are just part of the normal range. You should be using them for something special. Nothing should be annoying in that system.
It can cause confusion for stand in support people seeing .0 and/or .255 IP addresses, it's not usual.
Plus I split the network up, 254 addresses on 1 dhcp server and 254 on another. You can't exclude 0 & 255 from being assigned so I just reserve them so they don't get assigned.
I thought MS added failover DHPC servers in Server 2016 - so two servers could share a single range without risk of double assigning? Perhaps I misread something.
@https://www.itprotoday.com/windows-78/configuring-dhcp-failover-windows-server-2016 said:
DHCP failover is a new feature (available in Server 2012 and later versions) for ensuring high availability of DHCP server on an enterprise network. The two servers in a failover relationship share lease information including reservations, scope options, exclusion, policies, and filters
Looks like I was wrong - added in 2012.
I have dhcp running on 2 servers with non overlapping address scopes
server 1 hands out a.b.c.100 - 254
server 2 hands out a.b.d.100 - 254 -
@siringo said in What Are You Doing Right Now:
@Dashrender said in What Are You Doing Right Now:
@siringo said in What Are You Doing Right Now:
@scottalanmiller said in What Are You Doing Right Now:
@siringo said in What Are You Doing Right Now:
having to put in reservations for .255 & .0 ip addresses as windows dhcp issues them when you have a /23 scope. annoying.
Why would you be reserving them? In a /23 scope they are just part of the normal range. You should be using them for something special. Nothing should be annoying in that system.
It can cause confusion for stand in support people seeing .0 and/or .255 IP addresses, it's not usual.
Plus I split the network up, 254 addresses on 1 dhcp server and 254 on another. You can't exclude 0 & 255 from being assigned so I just reserve them so they don't get assigned.
I thought MS added failover DHPC servers in Server 2016 - so two servers could share a single range without risk of double assigning? Perhaps I misread something.
@https://www.itprotoday.com/windows-78/configuring-dhcp-failover-windows-server-2016 said:
DHCP failover is a new feature (available in Server 2012 and later versions) for ensuring high availability of DHCP server on an enterprise network. The two servers in a failover relationship share lease information including reservations, scope options, exclusion, policies, and filters
Looks like I was wrong - added in 2012.
I have dhcp running on 2 servers with non overlapping address scopes
server 1 hands out a.b.c.100 - 254
server 2 hands out a.b.d.100 - 254The scopes don't overlap, but the network does. How does a device know which one to ask for an address? Sounds like they currently just choose a range at random based on which DHCP server responds first? but... why?
-
@scottalanmiller said in What Are You Doing Right Now:
@siringo said in What Are You Doing Right Now:
@Dashrender said in What Are You Doing Right Now:
@siringo said in What Are You Doing Right Now:
@scottalanmiller said in What Are You Doing Right Now:
@siringo said in What Are You Doing Right Now:
having to put in reservations for .255 & .0 ip addresses as windows dhcp issues them when you have a /23 scope. annoying.
Why would you be reserving them? In a /23 scope they are just part of the normal range. You should be using them for something special. Nothing should be annoying in that system.
It can cause confusion for stand in support people seeing .0 and/or .255 IP addresses, it's not usual.
Plus I split the network up, 254 addresses on 1 dhcp server and 254 on another. You can't exclude 0 & 255 from being assigned so I just reserve them so they don't get assigned.
I thought MS added failover DHPC servers in Server 2016 - so two servers could share a single range without risk of double assigning? Perhaps I misread something.
@https://www.itprotoday.com/windows-78/configuring-dhcp-failover-windows-server-2016 said:
DHCP failover is a new feature (available in Server 2012 and later versions) for ensuring high availability of DHCP server on an enterprise network. The two servers in a failover relationship share lease information including reservations, scope options, exclusion, policies, and filters
Looks like I was wrong - added in 2012.
I have dhcp running on 2 servers with non overlapping address scopes
server 1 hands out a.b.c.100 - 254
server 2 hands out a.b.d.100 - 254The scopes don't overlap, but the network does. How does a device know which one to ask for an address? Sounds like they currently just choose a range at random based on which DHCP server responds first? but... why?
they do they just broadcast for an address, doesn't matter whether they're on network a or b, the scope options are the same for both scopes/servers. It's no different to having a single dhcp server that gives out addresses from a.b.c.100 - a.b.d.254
-
@siringo said in What Are You Doing Right Now:
@scottalanmiller said in What Are You Doing Right Now:
@siringo said in What Are You Doing Right Now:
@Dashrender said in What Are You Doing Right Now:
@siringo said in What Are You Doing Right Now:
@scottalanmiller said in What Are You Doing Right Now:
@siringo said in What Are You Doing Right Now:
having to put in reservations for .255 & .0 ip addresses as windows dhcp issues them when you have a /23 scope. annoying.
Why would you be reserving them? In a /23 scope they are just part of the normal range. You should be using them for something special. Nothing should be annoying in that system.
It can cause confusion for stand in support people seeing .0 and/or .255 IP addresses, it's not usual.
Plus I split the network up, 254 addresses on 1 dhcp server and 254 on another. You can't exclude 0 & 255 from being assigned so I just reserve them so they don't get assigned.
I thought MS added failover DHPC servers in Server 2016 - so two servers could share a single range without risk of double assigning? Perhaps I misread something.
@https://www.itprotoday.com/windows-78/configuring-dhcp-failover-windows-server-2016 said:
DHCP failover is a new feature (available in Server 2012 and later versions) for ensuring high availability of DHCP server on an enterprise network. The two servers in a failover relationship share lease information including reservations, scope options, exclusion, policies, and filters
Looks like I was wrong - added in 2012.
I have dhcp running on 2 servers with non overlapping address scopes
server 1 hands out a.b.c.100 - 254
server 2 hands out a.b.d.100 - 254The scopes don't overlap, but the network does. How does a device know which one to ask for an address? Sounds like they currently just choose a range at random based on which DHCP server responds first? but... why?
they do they just broadcast for an address, doesn't matter whether they're on network a or b, the scope options are the same for both scopes/servers. It's no different to having a single dhcp server that gives out addresses from a.b.c.100 - a.b.d.254
That's really weird. The problem with that setup is that you've got more stuff to maintain. You have the equipment of an HA setup, but without HA. You have complexity from the setup causing your techs to be confused (see your earlier comment) and if they (the DHCP servers) respond unevenly you could exhaust one pool and not the other. So it's negative in three ways without any positives. Your pool is overall smaller, the setup is overly complex, you have unnecessary risks, and you are paying for a full HA setup but not using it. Not to mention, you don't have a single source of reporting so looking up what is going on is really hard. Plus you have to configure two things instead of one.
Multiple DHCP servers on a single network is always considered a "no no". I've literally never heard of it done intentionally before. It's normally a mistake (by normally, I mean always.) And you turn off whatever one is not needed.
-
@siringo this setup is what we like to call "being weird." LOL What caused you to go down this path?
-
@scottalanmiller said in What Are You Doing Right Now:
@siringo this setup is what we like to call "being weird." LOL What caused you to go down this path?
it's what i learnt back in the day, (1990s).
hey this is good, i have a question.
i have noticed that i do get 1 server with most IPs handed out and the other with only a few.
if 1 server exhausts it's pool and receives a request for an address, does the server send back a 'sorry we're out of addresses' message?
coz if that's the case and PCs receive this message and don't re-broadcast their request, they could end up without an IP address.I'm not against changing my thinking, I just need learn better ways.
-
@siringo said in What Are You Doing Right Now:
it's what i learnt back in the day, (1990s).
Someone didn't understand DHCP basics and mistaught this. This is completely wrong from a general approach perspective, and all the certs that required this like Microsoft and Network+ talk about how this is a mistake and should be avoided. This is actually one of those "best practice" items that should always apply.
I can see what happened, most likely. Someone that was weak on networking was teaching this and got confused back then when Classful networking had recently existed and was still taught and they meant to tell you how to set up on VLANs and misspoke and it led to this. It's not hard to see how it could have happened accidentally. But rest assured, it's required for MS certs even back then that this is never how MS DHCP is supposed to be used (BIND would agree.)
Or maybe someone didn't understand HA and thought that they had come up with their own redundancy model back in a day when HA DHCP was rare or poorly known and taught their own made up architecture thinking that it would have some benefit. But it doesn't, especially not in the Microsoft world where those DHCP servers always offer HA.
But rest assured, they were wrong. It's bad (but obviously not show stopping) without any good parts. But it is super easy to fix.
-
@siringo said in What Are You Doing Right Now:
i have noticed that i do get 1 server with most IPs handed out and the other with only a few.
This is to be expected (and part of the core of why it's a bad design.) One will always have a speed advantage. Whether it is different hardware, not as busy, or just has shorter cables. One will tend to answer first and since it is "who responds first" rather than a round robin mechanism, one will almost always be the responder.
Oh, I just realized, likely it was someone who meant to be teaching DNS and instead did DHCP. DNS can be used in a similar way, but DHCP cannot. I bet whoever was teaching this got them switched in their minds and taught it wrong because of that. That can happen really easily.
-
@siringo said in What Are You Doing Right Now:
if 1 server exhausts it's pool and receives a request for an address, does the server send back a 'sorry we're out of addresses' message?
Yes, if Server 1 answers most of the time and runs out of addresses, it will deny an address to the new device and the second DHCP server will never get a chance to hand one out. That was the point I tried to make earlier about how you'd get far fewer IP addresses to use than you actually have because they'd be handed out asymmetrically.
-
I'm powering through Parks & Rec while my little one sleeps.
-
@siringo said in What Are You Doing Right Now:
coz if that's the case and PCs receive this message and don't re-broadcast their request, they could end up without an IP address.
Correct. So the setup only works reliably when you have less than ~250 total devices on the network. Once you hit as many devices as 1/2 of your total available, you start to risk not getting an address even though up to 1/2 of all possible addresses might still be available.
-
@DustinB3403 said in What Are You Doing Right Now:
I'm powering through Parks & Rec while my little one sleeps.
Good show, I really liked it.
-
@siringo said in What Are You Doing Right Now:
I have dhcp running on 2 servers with non overlapping address scopes
server 1 hands out a.b.c.100 - 254
server 2 hands out a.b.d.100 - 254So to go along with your question. This means each server has 155 addresses to hand out and there are 310 addresses in the pools.
With the way it is set up, only 155 addresses are guaranteed to be usable. But if you made the scope continuous not only would you get both sets, but you'd get c.100-254, but c.100-d.254 for 410 total addresses! Way more. Because you have loads being lost just because you are blocking off random addresses in the middle of your range, too.
-
Recording some videos for YouTube.
-
@EddieJennings said in What Are You Doing Right Now:
Recording some videos for YouTube.
They've already got quite a few.
-
This seems crazy, but CloudFlare mirrors the traffic numbers.
-
