| Author |
|
GadgetGuy
Super User
Joined: June 01 2008
Location: United States
Online Status: Offline
Posts: 942
|
| Posted: July 31 2013 at 15:02 | IP Logged
|
|
|
I have had a strange issue pop up in the last several
months, that keeps recurring about every 2 weeks, or so.
I cannot access the PH web server on my port 81. I have
other web servers running on ports 80 and 8080 that are
not affected.
But when I try to access my public IP address:81 (where
PH resides) I get a Page Not Found error.
I have disabled the Firewall and all virus protection,
but it makes no difference. The PH web server claims it
is "enabled" but looking at the Event Log there is
absolutely no evidence that a web request has ever
arrived.
The only way I have found to recover is to reboot the PH
PC.
I am at a loss to determine the root cause of this
behavior. The router port forwarding config appears
unchanged and correct.
Anyone ever have similar issues, or any ideas on what I
might try? 
__________________
Ken B - Live every day like it's your last. Eventually, you'll get it right!
|
| Back to Top |
|
| |
gg102
Senior Member
Joined: January 29 2013
Location: United States
Online Status: Offline
Posts: 246
|
| Posted: August 01 2013 at 20:41 | IP Logged
|
|
|
First, what changed over the past several months that now causes this issue?
Is it a NETWORK issue?
1) This almost sounds like a DHCP address lifetime/timeout/disconnect/release. Does this machine have a DHCP address or is it hard addressed?
2) At the moment that it is in failure mode, can you access the internet from that specific machine?
3) Are the other WEB servers you talked about running on the same machine?
4) When failing, can you assess the PH web page from another machine internally http://192.168.x.x:81?
5) While in failure mode can you ping/communicate with this machine from another or outside machine?
6) Could some other machine/server be intercepting the packets?
7) For test, can you isolate this machine on it's own net/subnet?
Is it a PH issue?
1) When it fails, can you restart PH and get the web server working or do you have to restart the machine?
2) When failing, can you access the PH WEB server on the SAME machine by http://127.0.0.1? If so, it may look more like a networking issue.
Is it a Router issue?
1) Do you have a service time configured in the router such that the service/port(81) is only allowed from x time until y time.
2) Have you tried to assign a specific DHCP address based on machine MAC? Some routers have this option. This could make port forwarding more reliable.
Is it a software issue?
1) Do you have something like Zone Alarm running? That might be turning off all networking on the machine.
2) Do you have any scheduled program running that trigger at some time? Things like a backup/imaging program etc...
I might be way off base but just some additional thoughts and background questions......
|
| Back to Top |
|
| |
dhoward
Admin Group
Joined: June 29 2001
Location: United States
Online Status: Offline
Posts: 4447
|
| Posted: August 01 2013 at 20:51 | IP Logged
|
|
|
GG,
WOW!...very thorough list of troubleshooting points. My guess is you do this kind of thing for a living  .
Dave.
|
| Back to Top |
|
| |
GadgetGuy
Super User
Joined: June 01 2008
Location: United States
Online Status: Offline
Posts: 942
|
| Posted: August 02 2013 at 06:43 | IP Logged
|
|
|
GG,
I second Dave. Very through and professional reply.
Good job!
I have done almost everything you suggest, but haven't
tried pinging that port, and a few others.
Will keep your list handy and the next time this happens,
will systematically try/test each item.
In the meantime, here are the answers (in red)
to the items I've tested already . . .
=====================
Is it a NETWORK issue?
1) This almost sounds like a DHCP address
lifetime/timeout/disconnect/release. Does this machine
have a DHCP address or is it hard addressed?
DHCP is not ever used to access this machine. It has a
STATIC IP address, which is always used ie,
192.168.1.90:81
2) At the moment that it is in failure mode, can you
access the internet from that specific machine?
Yes. the machine can reach out successfully to
all other computers on the LAN, as well as any Internet
web site.
3) Are the other WEB servers you talked about running on
the same machine? Yes. There are other web
servers running on the same machine. PRTG Network Monitor
on 80; and Blue Iris Camera controller on 8080
4) When failing, can you assess the PH web page from
another machine internally http://192.168.x.x:81?
No. That in point of fact is my issue! I
cannot access PH from my desktop computer, which I try to
do every morning to check that status of the automated
home environment and security
5) While in failure mode can you ping/communicate with
this machine from another or outside machine?
Haven't tried pinging, but am unable to access
PH from any inside/outside machine, but can access all
other web servers running on that machine.
6) Could some other machine/server be intercepting the
packets? Unlikely without major network
failure, also no other computers have experienced
significant changes and the problem is always fixed by
rebooting the PH computer
7) For test, can you isolate this machine on it's own
net/subnet? Not easily as this machine is the
home automation computer and everything depends on it.
If it was on a subnet then it would be essentially
unreachable. I'll put this one on hold unless a last
resort.
Is it a PH issue?
1) When it fails, can you restart PH and get the web
server working or do you have to restart the machine?
Restarting PH has no effect. Only rebooting
the computer cures the issue.
2) When failing, can you access the PH WEB server on the
SAME machine by http://127.0.0.1? If so, it may look more
like a networking issue. Have not tried
127.0.0.1. Always used 192.168.1.90:81 (will try on the
next failure event) When the failure occurs I cannot
access the PH web server even on the same machine, but I
can access other machines and web sites OK from this
machine.
Is it a Router issue?
1) Do you have a service time configured in the router
such that the service/port(81) is only allowed from x
time until y time. The router has no service
times set for any port
2) Have you tried to assign a specific DHCP address based
on machine MAC? Some routers have this option. This could
make port forwarding more reliable. Yes. Port
forwarding on the Cisco E2000 router is configured to the
target's MAC address.
Is it a software issue?
1) Do you have something like Zone Alarm running? That
might be turning off all networking on the machine.
There is no software running (that I am aware
of) that would affect networking. This machine is the
home automation center and is ONLY running four programs
(other than the standard Microsoft Win7 Professional load
like IE, etc) These programs are Allway Sync, PH, Blue
Iris, and LogTemp, a 1-wire sensor monitor app. These
have all been running for years without any issues until
now.
2) Do you have any scheduled program running that trigger
at some time? Things like a backup/imaging program etc...
Allway Sync handles automated backup as files
change, and PH periodically updates weather files, but
again these have been running for years without incident.
=====================
As you can see I've "been there, done that" for just
about everything, thus making this more of a mystery than
most issues. 
PS - One thing that may be relevant. For as long as I
can remember, when I try to access the PH web server from
the LAN or the Internet it often takes as long as 10
seconds before it appears in the browser window, while
the other two web servers running on that machine always
pop right up in a second or two. I always thought that
PH just was treating web access as a very low priority
task and took some time to "get around" to servicing the
request. Once the page is opened the subsequent
responses are as fast as normal. But the first access
effort has always been veeeeery slow. A clue, perhaps?
Edited by GadgetGuy - August 02 2013 at 06:53
__________________
Ken B - Live every day like it's your last. Eventually, you'll get it right!
|
| Back to Top |
|
| |
MrGibbage
Super User
Joined: October 23 2006
Location: United States
Online Status: Offline
Posts: 513
|
| Posted: August 02 2013 at 07:36 | IP Logged
|
|
|
Don't rule out a DHCP issue. Just because you use static
addressing on that machine, if your router tried to assign
that same IP address to another machine, you will
definitely have problems, with symptoms pretty exactly as
you describe. Periodic, seemingly random inaccessibility.
Check to see if you router can configure a DCHP range of
addresses and make sure you put all the static IP addresses
on your network outside of that range.
__________________
Skip
|
| Back to Top |
|
| |
dhoward
Admin Group
Joined: June 29 2001
Location: United States
Online Status: Offline
Posts: 4447
|
| Posted: August 02 2013 at 10:05 | IP Logged
|
|
|
Skip,
Good point. Always a possibility if you don't have your static IP's assigned outside the DHCP range. My favorite thing to do is to always use DHCP but assign the same IP based upon the MAC. I don't know of any routers that don't have this capability and its great because it gives you static IP's for the machines that need them without having to actually configure a static IP or worry about the DHCP range.
Ken,
Based upon your responses thus far, it seems to be a PowerHome issue more than anything else. The disturbing thing is that a restart of PH doesnt fix the issue and a reboot is required. This makes me wonder if its not a low level control such as the Catalyst socket control that PH uses for the webserver. A shutdown of PH (and make sure in task manager that it really shut down) and restart should clear out anything PH related so since this isnt working, Im suspecting something very low level in that socket communication chain.
You say that this issue started in the last several months. Do you recall if it was after a particular upgrade to PH or anything related to a change within PH? Since it seems to occur once every couple of weeks, it seems pretty regular. One thing to possibly try is saving off your plugins dir, pwrhome.db, and pwrhome.ini file, uninstalling PH, delete (or rename the directory) and reinstalling. After that, copying over your plugin dir, database, and INI file. This should rule out any underlying low level component corruption.
Dave.
|
| Back to Top |
|
| |
gg102
Senior Member
Joined: January 29 2013
Location: United States
Online Status: Offline
Posts: 246
|
| Posted: August 02 2013 at 10:21 | IP Logged
|
|
|
MrGibbage:
Good catch. I missed that one! ' '
GadgetGuy:
It seems that the next important thing to learn is if can you access the PH WEB server by http://127.0.0.1:81 when it fails. This will really help.
I would try that now, and see if it has a slow response when working.
( I feel like Columbo )
The answers to my Network 2 and 4 make MrGibbage's point more important as they appear to be in conflict with each other, while #5 tends to blow that theory since you can access other servers on the same machine.
Just one more thing if you don't mind. Are you accessing the other servers by name or address?
|
| Back to Top |
|
| |
GadgetGuy
Super User
Joined: June 01 2008
Location: United States
Online Status: Offline
Posts: 942
|
| Posted: August 02 2013 at 19:50 | IP Logged
|
|
|
Checked and verified that my router has DHCP enabled,
with Static Addresses reserved by MAC for key devices for
which I always want to know the IP Addr. All static
addresses are below xxx.xxx.xxx.99 and all DHCP address
are from 100 up.
So no issues there.
Dave, when I indicated that access fails every couple
weeks, I didn't mean to imply anything periodic. More
accurately I should have said a few times a month, on
average.
There have been no changes at all to PH. The only thing
different I can think of is the addition of my
findMyPhone App, which queries that ARP cache, but only
for the phone and never for the machine that PH runs on.
I can't imagine that is the culprit.
BUT, in the process of testing loopback and different
ports I realized that I access PH from my desktop
computer, not directly thru a http IP string in the
browser (as I do for ALL OTHER web services on that
machine), but via a short html page that opens, sizes,
and positions the window so it appears in the upper left
corner of my screen.
That html page contains the following . . .
Code:
<html>
<head>
<script language="JavaScript">
window.resizeTo(650,1150);
window.moveTo(100,100);
window.location = "http://192.168.1.90:81/ph-cgi/main";
</script>
</head>
<body>
<h1>If you see this, the redirect did not work properly.
</h1>
</body>
</html>
|
|
|
In testing it appears that this bit of code is what is
making PH take so long to open. Any ideas why this
causes such a delay?
I don't think it is related to the totally unresponsive
issue, however, as when that lockout occurs, I have
always tested using a http://192.168.1.90:81 string in
the browser.
The next time the lockout happens, I'll try the
127.0.0.1:81 approach as well as a total shutdown and
restart of PH. I've always just initialized, not
restarted so that will be an interesting test.
Thanks to all for your innovative and useful thoughts.
Good thinkin' all!
__________________
Ken B - Live every day like it's your last. Eventually, you'll get it right!
|
| Back to Top |
|
| |
GadgetGuy
Super User
Joined: June 01 2008
Location: United States
Online Status: Offline
Posts: 942
|
| Posted: August 03 2013 at 07:01 | IP Logged
|
|
|
OK - This is getting interesting! 
When I first attempted to access the PH web server this morning from my computer, it was not responding, so after about 7 seconds I exited and went to the PH computer and tried http://192.168.1.90:81 as well as http://127.0.0.1:81. Both attempts took 9-10 seconds before the web page appeared. I was watching the real-time Event Log window at the same time and the web access entry appeared just as the page popped up in the browser.
From the PH computer I then tried 192.168.1.90 ports 80 and 8080 (where other web servers are running) and in both cases their page appeared in ~1 second.
I then went back to my computer and tried the same experiment, and achieved virtually identical results.
From my computer, I then tried to access PH via my html redirect page. That took 11-12 seconds in multiple tries.
ALL OF THESE TESTS were done using Chrome, my default browser. On a whim I then tried to access the three different ports from both the PH and my computer, using IE 10 and in all cases all three web servers (including PH) popped the page into the browser in ~1 second. 
This would seem to implicate the Chrome browser, but I don't comprehend how it could be sensitive to port 81 unless there is some interaction.
I changed the PH web port to 8001 and tried the tests again with identical results to the above efforts. So it is not a port sensitive issue.
Apparently there is some interaction between Chrome and PH that dramatically slows initial accesses down. Once the page is accessed, all link clicks are normally responsive.
Any thoughts Dave? 
__________________
Ken B - Live every day like it's your last. Eventually, you'll get it right!
|
| Back to Top |
|
| |
dhoward
Admin Group
Joined: June 29 2001
Location: United States
Online Status: Offline
Posts: 4447
|
| Posted: August 03 2013 at 11:10 | IP Logged
|
|
|
Interesting. Certainly sounds like something between Chrome and PowerHome.
What I would do is go the eventlog and copy/paste the data from Chrome request along with a copy/paste from the IE request so we can see what differences there are in the request and headers. There may be something that Chrome is doing differently.
Dave.
|
| Back to Top |
|
| |
JANPOE
Newbie
Joined: August 03 2013
Location: United States
Online Status: Offline
Posts: 2
|
| Posted: August 03 2013 at 13:28 | IP Logged
|
|
|
I am having a similar issue with MYGARAGE. I can't access the webserver. I tried to resend an invite to my smartphone and I don't get an email back, I try to click on the link on an older AHPGateway email and i get a website not available message from iw.x10.com/garage?...
Any suggestions. Also I have tried to access the x10 forums and that comes up as a bad webpage as well...
|
| Back to Top |
|
| |
grif091
Super User
Joined: March 26 2008
Location: United States
Online Status: Offline
Posts: 1357
|
| Posted: August 03 2013 at 16:12 | IP Logged
|
|
|
The X10 forums, wiki, etc have been shut down. Go to the Smarthome forum and look for the information. One end user has started to host the X10 information on his own. There is also some information about another X10 site.
__________________
Lee G
|
| Back to Top |
|
| |
GadgetGuy
Super User
Joined: June 01 2008
Location: United States
Online Status: Offline
Posts: 942
|
| Posted: August 03 2013 at 16:35 | IP Logged
|
|
|
Thank Lee -
I was about to post a reply when yours popped up on my
screen. You beat me by 20 seconds. 
__________________
Ken B - Live every day like it's your last. Eventually, you'll get it right!
|
| Back to Top |
|
| |
JANPOE
Newbie
Joined: August 03 2013
Location: United States
Online Status: Offline
Posts: 2
|
| Posted: August 03 2013 at 16:38 | IP Logged
|
|
|
Thanks for the replies. .. are you familiar with the mygarage app. Does t
use my computer as a web server to operate the garage door or was
x10 hosting the webserver
|
| Back to Top |
|
| |
GadgetGuy
Super User
Joined: June 01 2008
Location: United States
Online Status: Offline
Posts: 942
|
| Posted: August 03 2013 at 16:50 | IP Logged
|
|
|
JANPOE-
Can't understand your question. Also your question has no
relationship to this posted thread. It is a totally
different issue. I suggest you open a new thread if you
want to pursue your question further.
Edited by GadgetGuy - August 03 2013 at 16:50
__________________
Ken B - Live every day like it's your last. Eventually, you'll get it right!
|
| Back to Top |
|
| |
GadgetGuy
Super User
Joined: June 01 2008
Location: United States
Online Status: Offline
Posts: 942
|
| Posted: August 03 2013 at 16:59 | IP Logged
|
|
|
dhoward wrote:
Interesting. Certainly sounds like
something between Chrome and PowerHome.
What I would do is go the eventlog and copy/paste the
data from Chrome request along with a copy/paste from the
IE request so we can see what differences there are in
the request and headers. There may be something that
Chrome is doing differently.
Dave.
|
|
|
Dave here are event Log captures of both Chrome and IE
accesses to my PH Home page.
The top one is IE and the bottom (under the red line) is
Chrome.
Woops! Maybe the image file upload isn't quite fixed?
Edited by dhoward - August 03 2013 at 17:29
|
| Back to Top |
|
| |
dhoward
Admin Group
Joined: June 29 2001
Location: United States
Online Status: Offline
Posts: 4447
|
| Posted: August 03 2013 at 17:31 | IP Logged
|
|
|
Ken,
Upload seems to have worked, the image code in the forum post appears to be what was messed up. I adjusted your post to have the proper image links for you.
Not sure why it happened. When I tested the image upload capability, I used FireFox. Let me know what browser you were using so I can test and make sure.
Dave.
|
| Back to Top |
|
| |
dhoward
Admin Group
Joined: June 29 2001
Location: United States
Online Status: Offline
Posts: 4447
|
| Posted: August 03 2013 at 17:56 | IP Logged
|
|
|
Ken,
Ive looked over the images and don't see anything at first glance that would lead me to believe there is a huge difference. One thing I did notice is that Chrome was using PH cookie authentication and IE didnt appear to use it. Im guessing that this is an internal machine that is on the trusted access list so Chrome really doesnt need to send the auth (not sure if this would cause a problem but I wouldnt think so).
I just tested Chrome on my system. The first request took 3 seconds (PH Device Status screen). I closed out Chrome, reopened and did the same request again and got the full page back in 1 second. My logs looked similar to yours including the cookie.
I next tested Firefox. This is the browser I normally use. On a fresh open and requesting the Device Status page, I got the results back in 1 second. Logs looked normal and had the cookie.
I then tried IE. Pretty much the same as Firefox and Chrome. Results in 1 second and the cookie was in the log.
Im not sure what would be different in your system. Do you have any plugins or extensions (not sure what Chrome calls them) for your Chrome browser? Im wondering if something there may be an issue.
Reviewing the PH webserver code, the entry is added to the eventlog as soon as the request is received. You said that you make the request from Chrome and the eventlog entry shows up about 9 seconds later (about the same time that the page paints). This almost sounds like Chrome is delaying the request but that doesnt make sense.
If you can think of anything unique or peculiar about your Chrome install, please let me know. If I can replicate it here, then it will be easier for me to try and fix.
Dave.
|
| Back to Top |
|
| |
dhoward
Admin Group
Joined: June 29 2001
Location: United States
Online Status: Offline
Posts: 4447
|
| Posted: August 03 2013 at 17:59 | IP Logged
|
|
|
Just a quick test of image upload using the Chrome browser.
It seems to have worked for me so Im not sure what went
wrong.
Dave.
|
| Back to Top |
|
| |
GadgetGuy
Super User
Joined: June 01 2008
Location: United States
Online Status: Offline
Posts: 942
|
| Posted: August 03 2013 at 19:44 | IP Logged
|
|
|
OK - I'll try again, this time with IE10.
Nope. IE wouldn't even upload the file. It keeps saying it must be a jpg/png/etcand smaller than 150Kb.
The file is a jpg that is only 94KB. Tried 3 times. All tries failed.
I'll email it to you!
__________________
Ken B - Live every day like it's your last. Eventually, you'll get it right!
|
| Back to Top |
|
| |
Active Topics 
Memberlist 
Search 
Help
Register 
Login
PowerHome Messageboard : PowerHome General
Topic: Unable to Access PH Web Server
