What Are You Doing Right Now
-
@EddieJennings said in What Are You Doing Right Now:
Looking at expanding the ephemeral port range for our production web server.
How does one do that? I mean... you can cut into the non-ephemeral range by a few. But that's risky, weird and gets you essentially nothing. You are really out of ports?
-
@scottalanmiller said in What Are You Doing Right Now:
@EddieJennings said in What Are You Doing Right Now:
Looking at expanding the ephemeral port range for our production web server.
How does one do that? I mean... you can cut into the non-ephemeral range by a few. But that's risky, weird and gets you essentially nothing. You are really out of ports?
https://technet.microsoft.com/en-us/library/cc735929(v=ws.10).aspx
This is what we've been trying to troubleshoot for a week and determine what's going on. The other alternative is to reduce the Time_Wait from the default of 240 seconds.
-
Playing with policies on a Watchguard
-
And trying not to fall asleep right now
-
@scottalanmiller Since these ports are being used for as source ports for outbound connections, where is the risk with making more ports available?
-
@EddieJennings said in What Are You Doing Right Now:
@scottalanmiller Since these ports are being used for as source ports for outbound connections, where is the risk with making more ports available?
The risk isn't huge, but the value is also... zero? Where do you plan to get more ports FROM?
-
@EddieJennings said in What Are You Doing Right Now:
This is what we've been trying to troubleshoot for a week and determine what's going on. The other alternative is to reduce the Time_Wait from the default of 240 seconds.
So several questions...
- Why do you have production web on Windows? Not that this is an / the issue, but... why?
- This is likely an application coding problem. This isn't something normally encountered by even the busiest websites.
- Are you properly reversed proxied? This isn't a problem you expect to see today.
- The logical alternative that everyone else does is increasing the size of the web farm.
-
@scottalanmiller By default Server 2012R2 uses 49,152 - 65,535. I could expand the range to 40,000 - 65,535.
-
Trying to resist the urge to smash a user's phone (or the user) with a hammer. He claims that the ringer volume doesn't work, and I asked him why he didn't just get a new phone (again, this happens at least twice a week). I sincerely hope he saw the anger, hatred, and frustration in my eyes, one of which was twitching from having to listen to his <long string of expletives deleted> stupid ass ring tone. He also has a laptop that makes windows sounds (I'd much rather be stabbed in the eye repeatedly with a rusty fish scaler than hear those crappy, over-digitized, horrible windows sounds).
-
@EddieJennings said in What Are You Doing Right Now:
@scottalanmiller By default Server 2012R2 uses 49,152 - 65,535. I could expand the range to 40,000 - 65,535.
Oh, that's weird. Yes, you can expand that for sure. That's still a lot to use, but you can grow quite a lot.
-
@scottalanmiller said in What Are You Doing Right Now:
- Why do you have production web on Windows? Not that this is an / the issue, but... why?
- This is likely an application coding problem. This isn't something normally encountered by even the busiest websites.
- Are you properly reversed proxied? This isn't a problem you expect to see today.
- The logical alternative that everyone else does is increasing the size of the web farm.
-
Environment is the way it's been forever (20+ years): The basic evolution, as I understand it, was an Access database that was eventually turned into an ASP.net site with SQL Server database (yes, I do know it's possible to use asp.net on Apache/Nginx).
-
I'm 100% sure it's a coding problem; however, right now, making either of these alterations would be a way to mitigate the problem until the code can be fixed.
-
No reverse-proxy exists. From what I've been able to gather, the issue is with port exhaustion talking to either (or all) our SQL server, REDIS server, or SMTP server. From what little I know of how a reverse proxy functions, I'm not sure how that helps.
-
I assume that means making more web servers for load balancing, yes?
-
@EddieJennings said in What Are You Doing Right Now:
- I assume that means making more web servers for load balancing, yes?
Correct
-
@EddieJennings said in What Are You Doing Right Now:
From what I've been able to gather, the issue is with port exhaustion talking to either (or all) our SQL server, REDIS server, or SMTP server.
Are these all external systems (different servers?) Normally SQL Server just needs a few ports, same with REDIS. SMTP might use a few but, how many emails can you send?
-
@scottalanmiller said in What Are You Doing Right Now:
@EddieJennings said in What Are You Doing Right Now:
From what I've been able to gather, the issue is with port exhaustion talking to either (or all) our SQL server, REDIS server, or SMTP server.
Are these all external systems (different servers?) Normally SQL Server just needs a few ports, same with REDIS. SMTP might use a few but, how many emails can you send?
I think the question is "how many emails can't we send?"
We live and die by E-mail. -
@EddieJennings said in What Are You Doing Right Now:
@scottalanmiller said in What Are You Doing Right Now:
@EddieJennings said in What Are You Doing Right Now:
From what I've been able to gather, the issue is with port exhaustion talking to either (or all) our SQL server, REDIS server, or SMTP server.
Are these all external systems (different servers?) Normally SQL Server just needs a few ports, same with REDIS. SMTP might use a few but, how many emails can you send?
I think the question is "how many emails can't we send?"
We live and die by E-mail.But... 2,500 per minute?
-
@scottalanmiller I don't know the answer to the question, but it's not 2500. Doing some quick math from looking at SendGrid we averaged 45 E-mails / minute (65k were sent yesterday).
-
@EddieJennings said in What Are You Doing Right Now:
@scottalanmiller I don't know the answer to the question, but it's not 2500. Doing some quick math from looking at SendGrid we averaged 45 E-mails / minute (65k were sent yesterday).
Right, that should not exhaust ports.
-
Sort and pitch - getting rid of the dust collectors around the office and office storage.
Finally found the iPod Classic 80GB - since we have iPads, iPhones and such.. not really working keeping... Better to let someone else have it who will use it. I can do more from everything else anyway.. who like a uni-tasker? Not I..
It's in near Mint condition, ... and I am sure I have the box for it too....
-
Just found this on Youtube thanks to Julian Ilett
Tagging @MattSpeller
Looks like it still needs some work... but could be a really neat tool.
-
Reminding myself to breathe.
