|
I've also had weird problems like that on my own machine too but I feel (possibly unreasonably) that I have more control when my machine is right here.
|
|
|
|
|
stheller2 wrote: but I feel (possibly unreasonably) that I have more control when my machine is right here.
If that's how you feel, let me tell you why you may not want that control and it really isn't about control anyway.
Maybe you'll change your mind about the cloud a little (you don't have to migrate all your projects, but your expectations may be more realistic).
First, imagine having nothing and now you want to run your own web application.
I have a few customers like that.
As a programmer, I can't install servers, but I can create an Azure account and run the web application in an App Service (€65, or about €50 if you only need the basics, or even €5/free if you need it really cheap!).
In fact, running in an App Service on Azure requires me to publish in Visual Studio, which is like three button clicks.
Or I can create (or automatically generate!) a pipeline in Azure DevOps, giving me the option to push my code to Git and have it released automatically.
You could even replace your App Service with an Azure Function (or Logic App) and get 10K executions for free, if it supports your use case.
I'll host a SQL database in Azure while I'm at it (€5 a month for a small one) and my customer has their own web application for about €70 a month.
For that you get daily back-ups, free TLS/SSL certificates, auto-scale, a bit of storage, staging slots (for A/B testing and/or deployment without downtime), the works.
You can, theoretically, run a whole lot of services for just that €70 a month.
If things get too crowded and your app becomes too slow, you can easily scale up.
You'll pay a bit extra each month, but it's a simple button click, rather than migrating an entire server or having downtime because you need to physically update your server.
You could compare those costs to those of a physical server, but remember, App Services are easy and just work.
A few button clicks and you'll have your own CDN (you probably wouldn't want a CDN on your own servers anyway).
No additional configuring required, no updates, no hassle.
These are by far the biggest costs, add a Storage Account, Key Vault, Service Bus, some Function Apps (for short, periodic or triggered code execution) and you'll pay short to nothing extra.
With a Logic App you can easily create workflows that trigger on incoming mail, Tweets, various events in M365, a whole lot, and then process that data and move it to some output (all clickable, your manager could do this!).
This is ideal for my smaller customers, but can just as well work for large customers.
Now take a high-availability scenario.
Using a cloud platform, you can build huge, redundant networks.
Think about mirroring your databases to another continent with automatic failover.
Your files are back-upped to multiple server racks in multiple locations on multiple continents.
This is, of course, all configurable, and the more copies the more you pay.
Using something like Docker, you can have multiple copies of your app running and automatically move your app to another server or even spin up new servers when one fails so your app is ALWAYS online.
Those scenario's are not possible with on-prem architecture for most organizations.
With Azure they become a possibility for even small organizations.
I haven't even started on (big) data possibilities yet.
You get an Azure Active Directory with it for free (that you can sync to your on-prem AD).
And even I, a programmer who knows little about hardware or networks, can make it happen!
So yeah, you give up some control, but that's the power of the cloud.
You could still spin up a VM in Azure and do all the work yourself, but that kind of defeats the purpose of the cloud.
Sometimes there's an outage (doesn't happen a lot, mind you), but your on-prem servers can fail you too!
When an application doesn't start that's often a real pain, but an application failing to start on your own computer is a pain too.
The cloud isn't some magic bullet that solves all your problems, but when people dismiss the cloud so easily they don't know what they talking about.
It's not about "having control", it's about creating business value that you couldn't possibly (or easily) achieve on your on-prem servers.
It's definitely more than just moving your IIS and/or SQL database to the cloud.
Happy to answer any questions you still might have 
|
|
|
|
|
I don't develop web applications, so most of that argument, however valid in other cases, doesn't apply to me.
By the same token, I don't run production jobs, just development, so scaling up for bigger production loads doesn't apply to me.
I'm also on a satellite internet connection with a data cap, which makes cloud computing very dubious anyway, especially with the size of test jobs I do run (billions to hundreds of billions of records, hundreds of gigabytes to terabytes of data).
Now what about my available hardware resources?
CPU/memory: My development server, purchased in 2019, has 2x Xeon Silver 4215 with 128 GB of DRAM; my test server, purchased last year, has 2x Xeon Gold 6326 with 256 GB of DRAM.
Storage: My development server has 2.5 TB of Optane SSD, 1 TB of NVMe flash SSD and 1 TB of SATA SSD; my test server has 3 TB of NVMe flash SSD with 1 TB of Optane SSD on the way here from an ebay purchase.
As for big data, my main home server has 2 TB of Optane DC persistent memory, which doesn't seem to be commercially available in the cloud at this time. My test server has 4 TB of Optane DC series 200 persistent memory borrowed from Intel; I'll buy some when it becomes available at a more reasonable price.
I own all of this equipment outright (other than the pmem loan from Intel).
So although I'm not disputing that cloud computing has its place, basically none of its benefits are relevant to my situation.
|
|
|
|
|
In that case it's not so much about "less control".
It also doesn't explain why you have cemented the resolve to never use cloud.
Anyway, in your situation I probably wouldn't use the cloud either.
Doesn't sound like a very good fit.
What do you do, if I may ask?
|
|
|
|
|
Sure. I'm the CTO of 2Misses Corp. We've just received US Letters Patent 11,254,590 for our foundational invention of a heterogeneous hash table and have half-a-dozen other patent applications in various stages of the patent process.
|
|
|
|
|
That's pretty cool and looks really complicated (math isn't my strong point).
It's also far from what I do in every day life (writing custom administrative software for various business clients).
The closest thing I do to what you're doing is probably:
var dict = new Dictionary<string, Whatever>();
employees.OrderBy(e => e.Name); 
|
|
|
|
|
Sure, that will do the same thing, just with slightly less scalability for billions of records. 
But really most of the complexity in the patent is the incredible detail needed to get a patent that might survive being attacked by big companies with lots of lawyers.
The basic idea is very simple: packing several variable-length records into a fixed-length bucket, so you don't need a pointer to every variable-length record but can instead keep many of those records directly in the table. This saves space (fewer pointers) and time (fewer random accesses to storage or memory).
|
|
|
|
|
I switched my web site to YAH (Yet Another Host) & the IP address changed, of course.
I can see the new IP address from everywhere, except my home.
Things I've done:
I checked the DNS propagation site & the new IP address is seen around the world (DNS Propagation Checker - Global DNS Testing Tool[^])
I made the change yesterday evening & finally gave up and went to bed because I couldn't see the new ip address only from my home.
Today I've:
I've flushed the local dns (100 times).
I've ran the release & renew ip address.
Turned off local computer & all network equipment.
I've turned off my Linksys wifi router for 20 minutes.
I've turned off my fiber cable modem.
I've backed up the wifi settings & complete restored them -- hoping to wipe the DNS cache on the wifi router -- I believe that is where this is coming from.
While the router is in a rebooting state, (with the cable modem turned off) if I ping I will see the old IP address.
ping newlibre.com
PING newlibre.com (64.209.142.205) 56(84) bytes of data.
From RADNet (192.168.5.1) icmp_seq=17 Destination Net Unreachable
At one point the router is completely down & I see:
$ ping newlibre.com
ping: newlibre.com: Temporary failure in name resolution
That I was thinking would be good.
Finally, everything comes back and again I ping the site & get the old 64.209.142.205 address.
Also, if you check the DNS that everyone sees it is the new one: DNS Propagation Checker - Global DNS Testing Tool[^]
This is insanity!
I can get to the new site from my work computer when I VPN to it & go to browser.
It's only here in my home that I cannot get to it. Insanity.
|
|
|
|
|
I literally dropped back to old wifi router firmware -- complete install of previous version.
It installed & rebooted & I flushed local dns & still gets old IP address.
Next, I installed the latest firmware on wifi router & rebooted again & flushed dns again.
Yep, still old IP address. I give up.
|
|
|
|
|
Geez,
I see why your old host was slow as molasses, they have 71 domains on your old VPS box.
Sometimes wonder how you guys that write books put your shoes on in the morning. Probably need a manual. ![Rose | [Rose]](https://codeproject.global.ssl.fastly.net/script/Forums/Images/rose.gif)
|
|
|
|
|
Randor, you are way smarter than me.
I've never thought anything different.
I just type. Most of the time things don't work.
So, I just type some more. Then, sometimes I get lucky.
|
|
|
|
|
Anyway,
You can query your DNS server and get the exact time the entry will expire. Just use nslookup and set the server to dytnohaa.metronetinc.net and query it.
You want the TTL value[^].
|
|
|
|
|
That’s cool. Was using s lookup but couldn’t quite figure it out. You’re much better than RTFM! 
|
|
|
|
|
Roger,
Your new/current hosting service has the TTL set to 86400 seconds. This is unusually high and also means that the next time your address changes you will have up to 24h downtime. I'd recommend lowering it.
|
|
|
|
|
Thanks, I set it to 3 hours. Hopefully that is realistic.
|
|
|
|
|
Yep, that's a good ttl value. 
But keep in mind that some ISP/networks will (rarely) ignore the TTL values and hold the cache for an arbitrary time regardless. Welcome to the internet, it's pure anarchy disguised as order.
|
|
|
|
|
Finally figured out how to get more details. Please help.
Why does this show 127.0.0.53 as the Server???
nslookup
> set debug
> newlibre.com
Server: 127.0.0.53
Address: 127.0.0.53#53
------------
QUESTIONS:
newlibre.com, type = A, class = IN
ANSWERS:
-> newlibre.com
internet address = 64.209.142.205
ttl = 1775
AUTHORITY RECORDS:
ADDITIONAL RECORDS:
------------
Non-authoritative answer:
Name: newlibre.com
Address: 64.209.142.205
------------
QUESTIONS:
newlibre.com, type = AAAA, class = IN
ANSWERS:
AUTHORITY RECORDS:
ADDITIONAL RECORDS:
------------
|
|
|
|
|
raddevus wrote: Why does this show 127.0.0.53 as the Server?
That's the server that you are querying. On Microsoft Windows this means that you are asking the local DNS Client (Dnscache) service. The Windows Dnscache service forwards the question upstream and caches the response.
|
|
|
|
|
I'm completely insane so I changed the Nameservers provided by the host at my google domain set up.
Previously I used customer nameservers (provided by host) and put them in at google.
I deleted those & added 1 record:
Host name
Type
TTL Data
newlibre.com.newlibre.com A 1 hour
162.246.23.194
I then saved it and immediately went to nslookup on my local machine and ran it and it INSTANTLY showed up:
Server: 127.0.0.53
Address: 127.0.0.53#53
Non-authoritative answer:
Name: newlibre.com
Address: 162.246.23.194
I'm completely insane now. Even though it now works.
|
|
|
|
|
Great,
You can query any public DNS service by changing the target server:
Try this:
nslookup
set debug
server 1.1.1.1
//Query your domain here 9.9.9.9 is Quad9
8.8.8.8 is Google
You can also target your router IP if it also runs a caching dns
|
|
|
|
|
Yep, I did that and when I targeted the new host's DNS I saw the new IP address.
When I would target metronet DNS it would always give me the old one.
But then when I added that A record in Google & switched to their default DNS servers then instantly my metronet DNS seemed to update & now I see the proper IP address too.
I'm still stumped. But that's my natural state. 
|
|
|
|
|
raddevus wrote: I'm still stumped. But that's my natural state.
You can debug this. Can you get the SOA record? The text you are pasting has "Non-authoritative answer" which means whatever DNS server you are asking doesn't own the record. Let's find the start-of-authority for your DNS zone.
In your nslookup use:
set type=soa
|
|
|
|
|
I wonder if you two keep this thread going like this if it will travel off this monitor and onto my second monitor.
sorry for the interruption, please continue.
|
|
|
|
|
I'm honestly and truly LOLing!!!
|
|
|
|
|
nslookup type=soa
Server: 8.8.8.8
Address: 8.8.8.8#53
** server can't find type=soa: NXDOMAIN
Earlier I went to a site which shows my DNS server:
Ended up being :
Country ISP DNS Server
US AS-CMN - Metronet 69.174.129.245
US AS-CMN - Metronet 2603:d010:b001:2::2
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
