I understand that the world is moving toward IPv6 because of IPv4 address exhaustion. I also understand that a server can listen for connections on both IPv4 and IPv6.
My question is, with respect to a web service: under what circumstances would a user connect using IPv6? Does it depend on their ISP?
I have a site where requesting a particular JSON file and finding it missing triggers a request for a fallback JSON file. However, rather than getting a 404 for the first file, I was getting the contents of index.html, which isn't valid JSON and created an error.
I traced the problem to this directive:
location / {
# First attempt to serve request as file, then
# as directory, then fall back to index.html
try_files $uri $uri/ /index.html;
}
How can I get it to serve a 404 for mysite.com/some/nonexistent/file.json, but still get index.html for mysite.com/?
I was briefly shown the networking setup for a rural hospital campus. Some of the things I was told or shown are:
This kind of networking is way beyond my experience level, such that I don't even know the terminology. I'm curious, though.
Basically I'm just trying to understand how this setup works and what the pieces do.
Thanks to anyone who will humor the naive questions of a programmer who lacks networking chops.
How can I back up and restore data from a Riak cluster? What commands do I need to run?
I have Riak installed on my machine and it's running (riak ping returns 'pong').
How can I check which version of Riak I have?
Nginx is said to be much more resource-efficient and easier to configure than Apache. A friend told me that "it can't do some things that Apache can do, but I don't need those things anyway."
Still, I'm curious: what kinds of things can Apache do that Nginx can't? I don't need an exhaustive list, just a general idea of scenarios where Apache would be a better choice.
I'm trying to track down a hit-or-miss bug in a web application. Sometimes a request completes just fine; sometimes it hangs and never finishes.
I see that Apache now has several requests listed on the server-status page as "sending reply," and that doesn't change. I'm testing on localhost, so there shouldn't ever be more than one.
Out of curiosity, I set MySQL to log all queries and I'm tail -fing the log file.
When things go OK, I see a pattern like this:
20 Connect root@localhost on dbname
20 Query (query #1)
20 Query (query #2)
(etc)
20 Quit
21 Connect (etc)
When it hangs, I see a pattern like this:
22 Connect root@localhost on dbname
22 Query (query #1)
//nothing happens, so I try the post again
23 Connect root@localhost on dbname
23 Query (query #1)
//nothing happens; try again
24 Connect (etc)
Here's my question: is MySQL logging attempted queries, or successful queries? In other words, if the last line I see is query #1, does that imply that query #1 or query #2 is hanging?
My guess is that the one I don't see is the problem, because the last one I see looks fine, but maybe the one I don't see is too screwed-up for MySQL to process. Thoughts?
I've got Apache installed on Windows. The folder name indicates that it's Apache 2.2, but I want to see the full version number. How can I do that?
I am setting up an Apache server (2.2.14) on Windows, and borrowing ideas from an existing XAMPP configuration. I am now trying to set up email functionality for PHP (5.3.1). (It needs to be able to handle attachments, just like the XAMPP server can.)
I've noticed that XAMPP comes with fake sendmail for Windows, but several things about the configuration don't make sense to me.
Here are a couple of lines from XAMPP's php.ini:
; For Unix only. You may supply arguments as well (default: "sendmail -t -i").
;sendmail_path = "C:\xampp\sendmail\sendmail.exe -t"
This is weird because:
sendmail_path is commented out by defaultI wondered if sendmail.exe is even necessary. So on my custom Apache server, which doesn't have sendmail.exe, I edited php.ini to add the correct information for SMTP, smtp_port, and sendmail_from. With that, it was able to send mail! But it's very slow.
Then I went back to the XAMPP server and renamed the sendmail directory, hoping to disable sendmail. That sort of worked - my PHP script claimed to have sent, but nothing happened.
So I'm very confused.
sendmail_path commented out, but apparently still uses the program it points to. How does it know?sendmail_path, but can somehow send mail without it. How does it do that?What do I actually need to do to configure the Apache server to send emails?
When serving a site with Apache, there are two ways that I know of placing some content at "servername/foo":
As an example of #2, XAMPP has these statements in a config file:
Alias /phpmyadmin "C:/xampp/phpMyAdmin/"
<Directory "C:/xampp/phpMyAdmin">
AllowOverride AuthConfig
Order allow,deny
Allow from all
</Directory>
...Which means "if the server is localhost, and somebody visits localhost/phpmyadmin, show them the contents of C:/xampp/phpMyAdmin, given the following permissions."
In this example, given that phpMyAdmin is accessible from the same URL, does it matter from a security perspective whether it's in the web root folder or just aliased to look as though it were?
(I realize there are similarly-titled questions, but this one is really 'how do I reset the password', and this one doesn't specify an OS and has only gotten Linux-oriented answers.)
On Windows XP, I just uninstalled MySQL and deleted the folder that it was installed in. I then rebooted and reinstalled MySQL.
When it comes back up, it still has the databases it had before the uninstall.
Where did it keep that database info? How do I completely remove MySQL and start over, as if I'd never installed it?
I work for a Sprint dealer. A customer of ours has sites where the only available internet connectivity is a Sprint cellular signal. They've been using a Cisco ASA5505 firewall. To give it WAN connectivity, they've been using a cheap Linksys router that has a PCMICA card slot, plugging in a Sprint card, and passing the connection via the Linksys LAN to the Cisco.
That solution is no longer available (the Linksys is going off the market), and frankly I thought it was awkward anyway. I'm trying to help them find a single Cisco device that will do everything for them. (Cisco is their preferred brand.)
I talked with Cisco and they suggested an 881G-S-K9.
The customer has said that they use a 10-user license and need support for 3 VLAN, whatever that means.
Sorry for any ignorance I may display by asking this question - I confess that I don't know a switch from a router from a firewall at this point in my career. The consumer-grade "router" I have at home may be all 3, for all I know. But hey, at least I'm trying to learn. :)
I have installed Apache 2.2.14 (Win32) on a Windows XP machine and am trying to add the latest PHP module. I downloaded the ZIP file from here labeled "VC9 x86 Non Thread Safe" and extracted to my Apache directory. I then copied php5.dll to Apache's bin directory and copied php.ini to C:\Windows.
In httpd.conf, I added these lines:
LoadModule php5_module "C:/Program Files/Apache Software Foundation/Apache2.2/bin/php5.dll"
AddType application/x-httpd-php .php
Now Apache will not start. error.log says this: "Can't locate API module structure php5_module in file C:/Program Files/Apache Software Foundation/Apache2.2/bin/php5.dll": No error"
I think I may have the wrong .dll file, because I found tutorials that use the filename php5apache2.dll and I didn't see that in the PHP package I got.
Also, I have seen references to a file called php5ts.dll, but I don't see that either.
What exactly do I need to make PHP5 work?
On a consumer-grade router, if I set a subnet mask to limit the number of concurrent users (for example, 255.255.255.248), could this cause problems for those users (accessing the web, etc)? Do I have to be aware of anything upstream at the ISP, or does this only affect my LAN?
This is the part I didn't get fully answered in my earlier question, in case it looks familiar. I did learn that:
(I'm not a network guy, so please humor me.)
I want to configure a typical consumer router as an open hotspot. However, the router's internet connection has limited bandwidth. Therefore, I'd like to limit how many people can connect to it simultaneously.
I could set DHCP to hand out a small number of addresses, but additional users could still connect if they specified their own IP address. So I'd rather set a hard limit.
I'm wondering if a subnet mask will do the trick. I (sort of) understand that subnet masks are a way of telling the network section of an IP address from the host part, and I think that means I can limit the number of possible addresses, and therefore users, on the router. So:
The list I was looking (sorry, I wasn't very clear) was this: if the first three octets are 255.255.255, what are the possible netmasks and how many hosts they support? I found this from the link Eddy provided. Answer:
Several of you have pointed out that limiting users is not the same as limiting bandwidth. It's a good point, but in this case, I'm just concerned with limiting the number of concurrent users.
So, my question again: if I simply assign one of the netmasks above to the router, will it restrict the number of possible users effectively, as listed above? Might it screw something else up?
From the Windows command prompt, I have FTPd to a Windows web server. I can get a file, and I can see a directory listing with dir, but I want to save that list locally.
I tried dir > c:\somefile.txt, and the file is created, but it's blank. Same thing if I do ls > c:\somefile.txt.
The result is the same when I FTP from a Linux box.
FTP sends back the following:
200 PORT command successful 150 Opening ASCII mode data connection for /bin/ls 226 Transfer complete
Our small company works with a large company and must access their customer data daily. In addition to assigning logins, they make us download a certificate on each computer we plan to use, and somehow set up a Cisco VPN inside Internet Explorer. Once logged in, that browser cannot access our corporate intranet, and if I visit whatismyip.com, it shows a different domain than the one at work.
What might be the reasoning for this strategy? If we are directly on their network rather than accessing limited info through a web interface, wouldn't that be less secure?
Joel Spolsky writes articles on Inc.com and regularly links to them from joelonsoftware.com. Whenever I try to read those articles from work (for example, this one), I see a blank page and the following in the title bar:
why-circuit-city-failed-and-why-bh-thrives.html (GIF image, 1x1 pixels)
Interestingly:
1) It doesn't matter which browser I use
2) I am able to read the articles from home.
Yes, I have brought this to the attention of the folks at Inc.com and even provided a traceroute, but haven't gotten an answer yet.
Now I'm just curious: what could cause a page to be "seen" as a 1x1 GIF?