|
/
Zope
/
Apsis
/
Pound Mailing List
/
Archive
/
2008
/
2008-11
/
URGENT: pound.socket in FreeBSD???
[
Re: [Pound Mailing List] reusing connections to ... ]
[
Service Unavailable. Please try again later. / ... ]
URGENT: pound.socket in FreeBSD???
"Human Servers" <serversrpeopletoo2(at)gawab.com> |
2008-11-08 08:37:06 |
[ SNIP ]
|
Hi everyone,
Sorry for the drama - but I have to know ASAP whether the FreeBSD install
of Pound uses a pound.socket or pound.sock file or neither.
If you've followed my issue so far:
1. Months ago I posted for a solution here but no one replied.
2. Now I URGENTLY need to get Pound going to launch a web app on another
server. (SAAS = $, livelihood..., my families future.)
3. The problem was caused by me temporarily disabling Pound to get SSH
going.
4. I now have to force Pound to start after boot.
5. Once started the listening address: 192.168.0.161:80 returns a 'Service
not available message'.
6. I can view the backend 192.168.0.160:8080 with/without Pound running.
7. I'm using FreeBSD. I have a pound.pid file but no pound.sock/(et). I
doubt I deleted/moved this file.
8. If pound.sock(et) isn't need on FreeBSD, I suspect I just need to be
shown how to run the poundctl command to re-enable the listerner.
PLEASE HELP ASAP!
:-(
-----------------------------------------------------------------------------------------------------------------------
Send big files for free. Simple steps. No registration.
Visit now http://www.nawelny.com
|
|
|
Re: [Pound Mailing List] URGENT: pound.socket in FreeBSD???
"Simon Matter" <simon.matter(at)invoca.ch> |
2008-11-08 10:57:37 |
[ SNIP ]
|
Hi,
I think one problem is that you really seem confused and your audience is
also confused.
> Hi everyone,
>
> Sorry for the drama - but I have to know ASAP whether the FreeBSD install
> of Pound uses a pound.socket or pound.sock file or neither.
>
> If you've followed my issue so far:
>
> 1. Months ago I posted for a solution here but no one replied.
> 2. Now I URGENTLY need to get Pound going to launch a web app on another
> server. (SAAS = $, livelihood..., my families future.)
> 3. The problem was caused by me temporarily disabling Pound to get SSH
> going.
I have absolutely no idea what relation those two have. So whatever you
did to 'get SSH going' may have broken your pound but I really can't think
why.
> 4. I now have to force Pound to start after boot.
> 5. Once started the listening address: 192.168.0.161:80 returns a 'Service
> not available message'.
> 6. I can view the backend 192.168.0.160:8080 with/without Pound running.
What and from where can you 'view'? That's an important question.
> 7. I'm using FreeBSD. I have a pound.pid file but no pound.sock/(et). I
> doubt I deleted/moved this file.
> 8. If pound.sock(et) isn't need on FreeBSD, I suspect I just need to be
> shown how to run the poundctl command to re-enable the listerner.
First, forget about the socket and about poundctl. The socket is a socket
and not a normal file and has no file content. It's being created once
pound starts up by pound itself. If you remove the corresponding config
line pound will still start and work but you just can't use poundctrl
(which is not needed for basic functioning of pound!).
So, I suggest you try to do things step by step. Maybe like this.
1) Log in to your pound server (I think it is 192.168.0.161, right?). Now
test the connection to your backend with
telnet 192.168.0.160 8080
If that works, try a 'GET /' and you should see some html garbage and
telnet should exit. If that works your backend is fine.
2) Start pound. Check that it is up and running with ps, then check that
it is listening on the correct host:port. On Linux this could be easily
done with
netstat -lpn | grep pound
The netstat command may need different args on FreeBSD but the manpage
should tell.
Then you see get something like this:
tcp 0 0 192.168.0.161:80 0.0.0.0:*
LISTEN 7808/pound
It clearly shows that process pound is listening on 192.168.0.161:80. If
it doesn't show up you may check with 'netstat -l' whether another process
already listens on 192.168.0.161:80.
3) If 1 and 2 are fine, it should only be a problem with your pound
config. Check that you get logs from pound when accessing 192.168.0.161:80
and try to find out why it fails if it still does.
One more suggestion: Many people told you what to check but you didn't
reply in detail. Maybe you first confirm 1 and 2 now and me or others may
help to get it going.
Simon
|
|
|
Re: [Pound Mailing List] URGENT: pound.socket in FreeBSD???
"Human Servers" <serversrpeopletoo2(at)gawab.com> |
2008-11-09 00:30:26 |
[ SNIP ]
|
Thanks Simon!
My original post was very clear as to my Pound problem:
http://www.apsis.ch/pound/pound_list/archive/2008/2008-06/1214199931000#1214199931000
and ended with a series of questions that I hoped would solve this problem
in a simple 'step-by-step' process.
Replies - like yours, tend to jump in at point when information previously
noted by me has been missed or ignored.
However, I appreciate ALL help, and in an effort to resolve this issue, I
have followed every suggestion in this mailing list and replied in detail
to each one, while trying to answer my original questions and adapt these
as I've learned more about Pound.
Thus, replying to you in detail:
Thank you for clearing up the socket creation process. This has removed a
large blackhole in my knowledge, and I feel a lot closer to solving this
problem now.
As I've written before, I can directly view the web page at backend
192.168.0.160:8080.
But I still telnetted to this as per your advice, added GET /, and yes I
got HTML back, which I recognise as being part of the web app.
I performed sockstat (FreeBSD equiv of netstat) and confirmed that Pound is
running and listening to 192.168.0.161:80 - as it should be.
So 1 and 2 are fine.
If you had read a previous mail from me, you would know that I have already
posted logs, and I have already noted that I haven't changed the config
file at all from when Pound originally worked.
var/log/messages:
server pound: no service "GET / HTTP/1.1" from 192.168.0.254 (the router)
And to reply to your bracketed point - I have NEVER believed that poundctl
is necessary for the normal running of Pound.
But to repeat myself (clealy) once agani:
I suspect I just need to be shown how to run the poundctl command to
RE-ENABLE the service.
My original sin was to disable Pound (thinking it has inhibiting SSH - as
again I've written earlier), which I must have done via the poundctl
command, and being a noob at Pound, I'm struggling to re-enable it.
I'm sorry if I seem rude, but I have been very clear in my replies and
questions, and repeating what I wrote only a few emails back seems rather
inefficient.
The reason for the length of my original post, was that I wanted to avoid
being misunderstood and repeating myself as I have thus done since.
I'm happy and appreciative of any and all help I get and I regard Pound as
a great free product.
-----------------------------------------------------------------------------------------------------------------------
Send big files for free. Simple steps. No registration.
Visit now http://www.nawelny.com
|
|
|
RE: [Pound Mailing List] URGENT: pound.socket in FreeBSD???
"Kiriki Delany" <kiriki(at)streamguys.com> |
2008-11-09 01:03:50 |
[ SNIP ]
|
I'm pretty sure that pound would restart its socket and all components,
whatever they may be, upon restarting the service
I never control pound with poundctl, just kill the PID and start pound
again.
I really think the nature of your issue, has to do with the jail. Can you
just remove pound from Jail for now?
If you get it to work with a wget from the pound server, then technically it
should work.
From following the posts, I'm not sure your having a Pound issue at all,
could just be a Jail issue
If it were me, I would just fire up pound in some alternate test configs,
outside of Jail, and you could also strip the HeadRequire settings.
i.e.
---------------- my pound.conf --------------------------
User "xxx"
Group "xxx"
ListenHTTP
Address 192.168.0.161
Port 80
LogLevel 2
# Virtual Host xxxxx.com
#
Service
BackEnd
Address 192.168.0.160
Port 8080
End
End
End
--------------- end of pound.conf ------------------------
Also, just another follow-up question, are you sure that there's no other
services running on the pound box on port 80? Like a web server for instance
at 192.168.0.161:80 ?
Thank you
-Kiriki
-----Original Message-----
From: Human Servers [mailto:serversrpeopletoo2(at)gawab.com]
Sent: Saturday, November 08, 2008 3:30 PM
To: pound(at)apsis.ch
Subject: Re: [Pound Mailing List] URGENT: pound.socket in FreeBSD???
Thanks Simon!
My original post was very clear as to my Pound problem:
http://www.apsis.ch/pound/pound_list/archive/2008/2008-06/1214199931000#1214
199931000
and ended with a series of questions that I hoped would solve this problem
in a simple 'step-by-step' process.
Replies - like yours, tend to jump in at point when information previously
noted by me has been missed or ignored.
However, I appreciate ALL help, and in an effort to resolve this issue, I
have followed every suggestion in this mailing list and replied in detail
to each one, while trying to answer my original questions and adapt these
as I've learned more about Pound.
Thus, replying to you in detail:
Thank you for clearing up the socket creation process. This has removed a
large blackhole in my knowledge, and I feel a lot closer to solving this
problem now.
As I've written before, I can directly view the web page at backend
192.168.0.160:8080.
But I still telnetted to this as per your advice, added GET /, and yes I
got HTML back, which I recognise as being part of the web app.
I performed sockstat (FreeBSD equiv of netstat) and confirmed that Pound is
running and listening to 192.168.0.161:80 - as it should be.
So 1 and 2 are fine.
If you had read a previous mail from me, you would know that I have already
posted logs, and I have already noted that I haven't changed the config
file at all from when Pound originally worked.
var/log/messages:
server pound: no service "GET / HTTP/1.1" from 192.168.0.254 (the router)
And to reply to your bracketed point - I have NEVER believed that poundctl
is necessary for the normal running of Pound.
But to repeat myself (clealy) once agani:
I suspect I just need to be shown how to run the poundctl command to
RE-ENABLE the service.
My original sin was to disable Pound (thinking it has inhibiting SSH - as
again I've written earlier), which I must have done via the poundctl
command, and being a noob at Pound, I'm struggling to re-enable it.
I'm sorry if I seem rude, but I have been very clear in my replies and
questions, and repeating what I wrote only a few emails back seems rather
inefficient.
The reason for the length of my original post, was that I wanted to avoid
being misunderstood and repeating myself as I have thus done since.
I'm happy and appreciative of any and all help I get and I regard Pound as
a great free product.
----------------------------------------------------------------------------
-------------------------------------------
Send big files for free. Simple steps. No registration.
Visit now http://www.nawelny.com
--
To unsubscribe send an email with subject unsubscribe to pound(at)apsis.ch.
Please contact roseg(at)apsis.ch for questions.
|
|
|
Re: [Pound Mailing List] URGENT: pound.socket in FreeBSD???
Paul Chvostek <paul+pound(at)it.ca> |
2008-11-09 19:27:00 |
[ SNIP ]
|
Hello, Human.
I haven't gone through every one of the messages in this thread, because
that would be an inefficient use of my time. But I am a long-time user
of Pound in FreeBSD, as am familiar with jails. I have found FreeBSD to
be an excellent platform on which to run Pound.
First off, this thing about poundctl... If you installed Pound from the
ports or a package system, as you should have, then your default
configuration includes the line:
Control "/var/run/pound"
and you should see something like this:
% ls -l /var/run/pound*
srwx------ 1 root wheel 0 Sep 29 05:52 /var/run/pound
-rw------- 1 root wheel 5 Sep 29 05:52 /var/run/pound.pid
Note the "s" at the beginning of the first line. Both files should show
the datestamp of the last time Pound was restarted, since they are both
recreated then.
I'll point out that I have never run Pound INSIDE a jail, because I
generally require a reverse proxy to sit between the Internet and a
private network. At this time, release versions of FreeBSD allow only a
single IP address per jail. This means that if your Pound jail has a
private jail IP (192.168.x.x), the outside world won't reach it. I
don't know what your network looks like; I just want to make sure you're
aware of this. If not, you may need to rearchitect things.
On Sat, Nov 08, 2008 at 11:30:26PM +0000, Human Servers wrote:
>
> My original post was very clear as to my Pound problem:
>
http://www.apsis.ch/pound/pound_list/archive/2008/2008-06/1214199931000#1214199931000
This post gives far too little information, and confuses your question
with unrelated assumptions. It's far better to re-ask a question
properly once you know what you're asking. Many of us will ignore a
question that looks like it involves hours of lessons. :)
> As I've written before, I can directly view the web page at backend
> 192.168.0.160:8080.
> But I still telnetted to this as per your advice, added GET /, and yes I
> got HTML back, which I recognise as being part of the web app.
So your web server is running. That's good.
> I performed sockstat (FreeBSD equiv of netstat) and confirmed that Pound is
> running and listening to 192.168.0.161:80 - as it should be.
Sockstat is not an equivalent of netstat. It provides different
information. In this case, it provides more useful information, so it's
the right tool to use. Users of other operating systems (Linux,
Solaris, etc) can get some of the information that sockstat provides
with "netstat -an | grep -w LISTEN"
> If you had read a previous mail from me, you would know that I have already
Don't be rude. As I said, your initial query was full of holes. I
wouldn't have read past the first two blatantly erroneous assumptions,
and certainly would not have answered it.
> I suspect I just need to be shown how to run the poundctl command to
> RE-ENABLE the service.
You can always re-enable pound by restarting the daemon:
% sudo /usr/local/etc/rc.d/pound restart
None of the state that you affect using poundctl will withstand the
service being killed and restarted.
> I'm sorry if I seem rude, but I have been very clear in my replies and
> questions, and repeating what I wrote only a few emails back seems rather
> inefficient.
Far more inefficient is forcing everyone else to re-read your old emails
and try to figure out what your actual problem is. It would be far
better for us to get you to the point that you actually understand what
is going on, so that you can solve the problem yourself. Then, as a
knowledgeable user, you can pass on the favour to the next person who
emails the list about his socks, jails and the phase of the moon.
--
Paul Chvostek <paul(at)it.ca>
Operations / Abuse / Whatever
it.canada, hosting and development http://www.it.ca/
|
|
|
Re: [Pound Mailing List] URGENT: pound.socket in FreeBSD???
"Human Servers" <serversrpeopletoo2(at)gawab.com> |
2008-11-10 01:41:12 |
[ SNIP ]
|
Thanks Paul!
I wasn't intending to be rude, nor would I expect someone to read everyone
of my emails.
But it was rude for others to accuse me of incorrectly assuming things
without first searching my previous mails to see if I had made the
assumption or not. I felt I should point out that I have already provided
information to the contrary.
In the main most replies haven't made ill-informed assumptions and have
been very helpful.
My questions didn't beg 'hours of lessons', and were rather direct, and
have evolved as I have tried the suggestions of those who have replied.
If I were to follow your advice: 'It's far better to re-ask a question
properly once you know what you're asking', then I doubt anyone would ask
questions.
I also tried to provide enough information in my first post so anyone could
correct my assumptions and questions. And I did so under rather logical
headings - trying to make it easy to read and see that I wasn't making to
many assumptions.
I provided server info, jails info, process info, my config, etc.
it wasn't perfect information - but it was a better start than many other
mails
I have read here.
But you can blame me, the pound noob, or the inefficiencies of mailing
lists/online communications.
There appear to be a lot of mails that get lost/unanswered in this this
mailing list, and the mailers ask more erudite questions than I have.
Perhaps the knowledge level for entry has been set too high on this mailing
list?
A community wiki/google group combination may be far more efficient,
accessible, and friendlier.
This aside, I have aimed to learn a lot here and contribute back in the
future!
Thank you for taking the time to read my initial mail and take the time to
reply.
Back to being productive then!
>and you should see something like this:
>
> % ls -l /var/run/pound*
> srwx------ 1 root wheel 0 Sep 29 05:52 /var/run/pound
> -rw------- 1 root wheel 5 Sep 29 05:52 /var/run/pound.pid
After forcing pound to start, I do see this exactly (with a different time
of course).
>I'll point out that I have never run Pound INSIDE a jail, because I
>generally require a reverse proxy to sit between the Internet and a
>private network. At this time, release versions of FreeBSD allow only a
>single IP address per jail. This means that if your Pound jail has a
>private jail IP (192.168.x.x), the outside world won't reach it. I
>don't know what your network looks like; I just want to make sure you're
>aware of this. If not, you may need to rearchitect things.
Thanks, I am aware of the IP limitations of jails, which is why I am using
them.
They have worked fine to date with a firewall and NAT/port-forwarding, so
there is no need to rearchitect at this time.
As I have previously noted, I have not messed up anything other than pound,
so the jails are not an issue.
>> As I've written before, I can directly view the web page at backend
>> 192.168.0.160:8080.
>> But I still telnetted to this as per your advice, added GET /, and yes I
>> got HTML back, which I recognise as being part of the web app.
>
>So your web server is running. That's good.
Yes thanks, I have noted many times here that Apache is running fine, and
the only issue appears to be Pound.
>> I performed sockstat (FreeBSD equiv of netstat) and confirmed that Pound
>is
>> running and listening to 192.168.0.161:80 - as it should be.
>
>Sockstat is not an equivalent of netstat. It provides different
>information. In this case, it provides more useful information, so it's
>the right tool to use. Users of other operating systems (Linux,
>Solaris, etc) can get some of the information that sockstat provides
>with "netstat -an | grep -w LISTEN"
My bad, I lumped them together for Linux users re finding sockets in
FreeBSD.
>> If you had read a previous mail from me, you would know that I have
>already
>
>Don't be rude. As I said, your initial query was full of holes. I
>wouldn't have read past the first two blatantly erroneous assumptions,
>and certainly would not have answered it.
Anyone reading my third line would see that I note my assumptions were
wrong:
'SSH it seems is fine regardless of proxy, and now I can't find
my note of the command I used to disable Pound, so that I may
reverse this.'
It would be rude to dismiss an email because of the first two assumptions,
particularly when they are knocked over in the next line.
My only assumption was that pound was preventing SSH from working.
And I noted that after I realised it wasn't, I was seeking to re-enable
Pound.
I included the information re SSH to allow someone to correct me that Pound
may have caused a problem with SSH - if this were possible.
I divided my mail into logical topcial headers, and tried more than most to
make it a quick easy read.
I can't account for readers temperaments nor their attentions spans.
I provided all the information I could to avoid repeating myself and to
prevent wasting others time asking me obvious questions.
>> I suspect I just need to be shown how to run the poundctl command to
>> RE-ENABLE the service.
>
>You can always re-enable pound by restarting the daemon:
>
> % sudo /usr/local/etc/rc.d/pound restart
>
>None of the state that you affect using poundctl will withstand the
>service being killed and restarted.
Thanks, that's good to know, and news to me.
I'm still trying to work out how to use poundctl to enable/disable
different services.
For now sudo /usr/local/etc/rc.d/pound restart is sufficient to right
things.
I'd like to ask you some questions about Pound on FreeBSD, which I'd hope
would resolve this problem, but I there seems little point.
Thanks
-----------------------------------------------------------------------------------------------------------------------
Send big files for free. Simple steps. No registration.
Visit now http://www.nawelny.com
|
|
|
RE: [Pound Mailing List] URGENT: pound.socket in FreeBSD???
"Human Servers" <serversrpeopletoo2(at)gawab.com> |
2008-11-10 02:07:37 |
[ SNIP ]
|
Thanks Kiriki!
Tbh, I really don't want to risk breaking anything else by moving pound.
But I did follow your advice to remove HeadRequires, which initially got
the
site back up - internally. That is, 192.168.0.161:80 was being directed by
Pound to 192.168.0.160:8080.
Then I added HeadRequires back, restarted Pound, and oddly, the site was
again visible -
internally.
But after restarting Pound after a reboot, internally I get the 'Service
unavailable' message.
So I removed the HeadRequires, but this time the site was not internally
visible - with no 'Service unavailable message' too.
Even when the site was temporarily visible via 192.168.0.161, externally
visitors were seeing the message:
'The server at xxxxxxx is taking too long to respond...'.
Apache is configured to listen to the backend address: 192.168.0.160:8080 -
not 161:80, so no interference there.
Thanks for your help!
I'll keep plodding on!
-----------------------------------------------------------------------------------------------------------------------
Send big files for free. Simple steps. No registration.
Visit now http://www.nawelny.com
|
|
|
RE: [Pound Mailing List] URGENT: pound.socket in FreeBSD???
"Human Servers" <serversrpeopletoo2(at)gawab.com> |
2008-11-10 03:12:40 |
[ SNIP ]
|
I may have found the problem:
After forcing Pound to start:
/usr/jails/proxy.xxxxxxx.com/usr/local/sbin/pound -f
/usr/jails/proxy.xxxxxxx.com/usr/local/etc/pound.conf
ls -l /var/run/pound* shows
srwx------ 1 root wheel 0 Nov 10 15:09 /var/run/pound
-rw------- 1 root wheel 4 Nov 10 15:09 /var/run/pound.pid
but 'The service is not available. Please try again later.' message appears
internally and externally.
If I try to restart Pound with:
/usr/jails/proxy.xxxxxxx.com/usr/local/etc/rc.d/pound restart
I get this error:
'can't open configuration file "/usr/local/etc/pound.cfg" (No such file or
directory) - aborted'
Please correct me if I'm wrong (I did ask this questions a few posts back):
On Linux is the config file called pound.cfg and on FreeBSD pound.conf?
And if it is pound.conf, then where do I update where Pound looks for the
config file?
Thanks!
-----------------------------------------------------------------------------------------------------------------------
Send big files for free. Simple steps. No registration.
Visit now http://www.nawelny.com
|
|
|
RE: [Pound Mailing List] URGENT: pound.socket in FreeBSD???
"Kiriki Delany" <kiriki(at)streamguys.com> |
2008-11-10 03:24:11 |
[ SNIP ]
|
If you don't explicitly define the config file, then it looks for it in the
default location /usr/local/etc/pound.cfg
Your defining it when you start it with
/usr/jails/proxy.xxxxxxx.com/usr/local/sbin/pound -f
/usr/jails/proxy.xxxxxxx.com/usr/local/etc/pound.conf
Running this command /usr/jails/proxy.xxxxxxx.com/usr/local/etc/rc.d/pound
restart
Should be equivalent to just running
/usr/jails/proxy.xxxxxxx.com/usr/local/sbin/pound
Try to figure out what's causing this behavior, it should be consistent
"But I did follow your advice to remove HeadRequires, which initially got
the site back up - internally. That is, 192.168.0.161:80 was being directed
by Pound to 192.168.0.160:8080.
Then I added HeadRequires back, restarted Pound, and oddly, the site was
again visible - internally."
-Kiriki
-----Original Message-----
From: Human Servers [mailto:serversrpeopletoo2(at)gawab.com]
Sent: Sunday, November 09, 2008 6:13 PM
To: pound(at)apsis.ch
Subject: RE: [Pound Mailing List] URGENT: pound.socket in FreeBSD???
I may have found the problem:
After forcing Pound to start:
/usr/jails/proxy.xxxxxxx.com/usr/local/sbin/pound -f
/usr/jails/proxy.xxxxxxx.com/usr/local/etc/pound.conf
ls -l /var/run/pound* shows
srwx------ 1 root wheel 0 Nov 10 15:09 /var/run/pound
-rw------- 1 root wheel 4 Nov 10 15:09 /var/run/pound.pid
but 'The service is not available. Please try again later.' message appears
internally and externally.
If I try to restart Pound with:
/usr/jails/proxy.xxxxxxx.com/usr/local/etc/rc.d/pound restart
I get this error:
'can't open configuration file "/usr/local/etc/pound.cfg" (No such file or
directory) - aborted'
Please correct me if I'm wrong (I did ask this questions a few posts back):
On Linux is the config file called pound.cfg and on FreeBSD pound.conf?
And if it is pound.conf, then where do I update where Pound looks for the
config file?
Thanks!
----------------------------------------------------------------------------
-------------------------------------------
Send big files for free. Simple steps. No registration.
Visit now http://www.nawelny.com
--
To unsubscribe send an email with subject unsubscribe to pound(at)apsis.ch.
Please contact roseg(at)apsis.ch for questions.
|
|
|
RE: [Pound Mailing List] URGENT: pound.socket in FreeBSD???
SBR <sbr(at)rlfans.com> |
2008-11-10 20:49:50 |
[ SNIP ]
|
Hi,
Just to confirm when you say "192.168.0.161:80" is/is not directed to the
backend do you mean requesting "http://192.168.0.161/" is/is not directed?
Or do you mean that requesting "http://www.xxxxxx.com/" where
www.xxxxxx.com resolves to 192.168.0.161 is/is not directed?
Oh, and have you tried "poundctl -c /var/run/pound"?
Regards,
Andrew
|
|
|
RE: [Pound Mailing List] URGENT: pound.socket in FreeBSD???
"Human Servers" <serversrpeopletoo2(at)gawab.com> |
2008-11-11 00:06:06 |
[ SNIP ]
|
Thanks Andrew!
Our firewall is set up so a public static IP is answered by the jail/Pound
of 192.168.0.161:80. (NAT/Portforwarding is used.)
Before I broke Pound, Pound would direct the 192.168.0.161:80 request to
the backend of 192.168.0.160:8080, which is the www.xxxxxxxxxxx.com
address.
Please forgive my sloppy description of how Firewall/Jail/Pound does this.
It works well. I've only messed wiith Pound.
192.168.0.161:80 and www.xxxxxxxxxxx.com both return the message:
'Theservice is not available. Please try again later.'
Directly viewed, the backend 192.168.0.160:8080 resolves to the correct
html page,
which indicates Apache is listening fine as per usual.
/usr/jails/proxy.xxxxxxxxxxx.com/usr/local/sbin/poundctl -c /var/run/pound:
0. http Listener 192.168.0.161:80 a
0. Service active (5)
0. Backend PF_INET 192.168.0.160:8080 active (5 0.000 sec) alive
-1. Global services
If the output above is wrong for what should be listed - please suggest how
to correct.
Thanks!!!
-----------------------------------------------------------------------------------------------------------------------
Send big files for free. Simple steps. No registration.
Visit now http://www.nawelny.com
|
|
|
RE: [Pound Mailing List] URGENT: pound.socket in FreeBSD???
"Human Servers" <serversrpeopletoo2(at)gawab.com> |
2008-11-11 00:13:29 |
[ SNIP ]
|
Thanks Kiriki!
>Your defining it when you start it with
>
>/usr/jails/proxy.xxxxxxx.com/usr/local/sbin/pound -f
>/usr/jails/proxy.xxxxxxx.com/usr/local/etc/pound.conf
>
>Running this command /usr/jails/proxy.xxxxxxx.com/usr/local/etc/rc.d/pound
>restart
>
>Should be equivalent to just running
>/usr/jails/proxy.xxxxxxx.com/usr/local/sbin/pound
That adds to what Paul wrote and really has cleared up some of my
confusion!
Thanks!
>Try to figure out what's causing this behavior, it should be consistent
Hehe, yes it should be. Hopefully today is the day I get this fixed and
consistent!
I *really* appreciate all the help everyone is giving here!!
Once the dust settles, I'd like to start a Pound wiki if that's cool?
Thanks!!!
-----------------------------------------------------------------------------------------------------------------------
Send big files for free. Simple steps. No registration.
Visit now http://www.nawelny.com
|
|
|
RE: [Pound Mailing List] URGENT: pound.socket in FreeBSD???
SBR <sbr(at)rlfans.com> |
2008-11-11 00:59:24 |
[ SNIP ]
|
Hi,
The output from poundctl looks right.
To get the 'The service is not available. Please try again later' either
your request does not match any service or pound thinks all the backends
for the matching service are down. As pound is showing the backend as
alive it looks like the HeadRequire directive is causing the problem.
I suggest removing that line. It isn't required for what you want to do to
work.
Regards,
Andrew
|
|
|
RE: [Pound Mailing List] URGENT: pound.socket in FreeBSD???
"Human Servers" <serversrpeopletoo2(at)gawab.com> |
2008-11-11 02:33:18 |
[ SNIP ]
|
Thanks Andrew!
I removed HeadRequire again (Kiriki had also suggested this), and now
www.xxxxxxxxx.com works - internally.
But externally, www.xxxxxxxxx.com is timing out with same message 'A
network error has occurred...'
I had the same 'good luck' after following Kiriki's advice until I rebooted
the server and forced started Pound again - then I lost the internal view
again.
I hope the same doesn't happen next time I reboot! :D
I guess the problem must the firewall (which is admin'd by someone else)?
But I'll have to have a look at Apache's httpd.conf too!
Thanks!
-----------------------------------------------------------------------------------------------------------------------
Send big files for free. Simple steps. No registration.
Visit now http://www.nawelny.com
|
|
|
|