Re: apache2's handling of IP version 6
Hi, Tom H All that you really need for ipv6 is the ::1 ... line. Having one line less on your X-less box won't make a difference. Is this your entire /etc/hosts? Don;t you have any ipv4 settings?! The entire text is the following: 127.0.0.1 localhost 192.168.xx.xx .xxx.xx Is the 192.168... line added by network-manager? The 192.168... line exists from the beginning after I installed Squeeze by debian-6.0.3-i386-netinst.iso (The package of network-manager is not installed. ) Thanks, Satoru -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121010022030.1c93f7e8.trn...@otip.jp
Re: apache2's handling of IP version 6
On Sat, Oct 6, 2012 at 3:01 PM, Satoru Otsubo trn...@otip.jp wrote: Does ipv6 come up cleanly if you do a soft or hard restart of apache in rc.local? I can't understand a soft or hard restart. soft: service apache2 relaod service apache2 graceful hard: service apache2 restart All that you really need for ipv6 is the ::1 ... line. Having one line less on your X-less box won't make a difference. Is this your entire /etc/hosts? Don;t you have any ipv4 settings?! The entire text is the following: 127.0.0.1 localhost 192.168.xx.xx .xxx.xx Is the 192.168... line added by network-manager? -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CAOdo=sw5nxezowxfkr654yaaqrgg5mgbkdkqa5cupb9lqjs...@mail.gmail.com
Re: apache2's handling of IP version 6
On Fri, Oct 5, 2012 at 6:54 AM, Stan Hoeppner s...@hardwarefreak.com wrote: On 10/4/2012 12:06 PM, Tom H wrote: On Thu, Oct 4, 2012 at 7:23 AM, Stan Hoeppner s...@hardwarefreak.com wrote: On 10/4/2012 3:46 AM, Rick Thomas wrote: On Oct 3, 2012, at 8:40 PM, Satoru Otsubo wrote: But the phenomena are same, that is, When booting my PC, apache2 failed to start. And when I executed the following: # /etc/init.d/apache2 restart apache2 started successfully with the dual stack. Why this phenomena happens ? Is the apache2 daemon starting before the ipv6 part of the network configuration is completely up? His log error does seem to indicate Apache is starting before the ipv6 stack is available--one of the downsides of parallel init. Where did you see the log? The OP pasted this in his 2nd post: Starting web server: apache2[Thu Oct 04 01:02:03 2012]: [crit] (EAI 9)Address family for hostname not supported: Alloc_listener: failed to set up sockaddr for ::: Syntax error on line 2 of /etc/apache2/ports.conf Listen setup failed. Action 'start' failed. The Apache error log may have more information. failed! Thanks. I'd checked the archive before emailing you and missed it - for the second time! I think Chris Bannister is onto something. I previously overlooked the syntax error msg. A syntax error would explain the socket setup failure. Probably something as simple as a typo in ports.conf. I don't understand why there's a syntax error at boot but not when apache's restarted but we'll see... As I pointed out in an earlier post, Listen 80 will enable both ipv4 and ipv6, and remove the possibility that there's a typo in the ip address(es). Since Debian (and all distributions that I know) compile apache with an option to map v4 to v6, running ss -ntl will show just one ipv6 socket. Maybe having both Listen *:80 and Listen [::]:80 will create two sockets, if that's what the OP's after (or one of the things that he's after). I can't check because I don't have ipv6 and I'm not sure that apache'll interact the same way with an ipv6 link-local address as it would with a real ipv6 address. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CAOdo=SztX2Ye4edz8arZZEmqepbo+-x=e7vxdbyap3pzvu_...@mail.gmail.com
Re: apache2's handling of IP version 6
On Fri, Oct 5, 2012 at 12:42 PM, Satoru Otsubo trn...@otip.jp wrote: On Thu, Oct 04, 2012 at 04:53:23AM +0900, Satoru Otsubo wrote: By the way, I thought the following concerning the problem in my CGI PC (my server PC): Internet-Router--Lan-my PC When booting PC, apache2 decides that ipv6 is Ok if it gets the ipv6's information and ipv6 is not Ok if it can't get the ipv6's information. And apache2 gets the ipv6's information from the router's RA(Router Advertisements). That explains the at-boot v/s after-restart behavior... Can be started later, based on a sysvinit stage like $network, etc. Does ipv6 come up cleanly if you do a soft or hard restart of apache in rc.local? Then after booting completed, I executed, # netstat -an And I found there exists Listen 80 and Listen 443 in the tcp6 lines. My thought is wrong ? By the way, on the PC of Squeeze desktop install (my second PC), After booting completed, on command of # netstat -an , I can always find Listen 80 and Listen 443 in the tcp6 lines. (Without # /etc/init.d/apache2 restart) It's normal; apache listens on the ipv6 socket for both ipv4 and ipv6. In fact, I found, /etc/hosts of my GUI PC includes # The following lines are desirable for IPv6 capable hosts ::1 localhost ip6-localhost ip6-loopback fe00::0 ip6-localnet ff00::0 ip6-mcastprefix ff02::1 ip6-allnodes ff02::2 ip6-allrouters ff02::3 ip6-allhosts whereas, /etc/hosts of my CUI PC only includes # The following lines are desirable for IPv6 capable hosts ::1 ip6-localhost ip6-loopback fe00::0 ip6-localnet ff00::0 ip6-mcastprefix ff02::1 ip6-allnodes ff02::2 ip6-allrouters All that you really need for ipv6 is the ::1 ... line. Having one line less on your X-less box won't make a difference. Is this your entire /etc/hosts? Don;t you have any ipv4 settings?! -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CAOdo=sz5n8zfoe6k51g-qqz7pig1q8o1kc0rdsohyn4bmac...@mail.gmail.com
Re: apache2's handling of IP version 6
Hi,
Tom H
Does ipv6 come up cleanly if you do a soft or hard restart of apache
in rc.local?
I can't understand a soft or hard restart.
But I examined as follows:
..Case 1..
I editted /etc/rc.local as follows: (Do I have to add exit 0 at end ?)
#!/bin/sh -e
#
# rc.local
#
# This script is executed at the end of each multiuser runlevel.
# Make sure that the script will exit 0 on success or any other
# value on error.
#
# In order to enable or disable this script just change the execution
# bits.
#
# By default this script does nothing.
/etc/init.d/apache2 restart
exit 0
(Reference)
# /etc/rc2.d/
README S18apache2 S19openbsd-inetd S21bootlogs
S14portmap S19acpidS19spamassassin S22rc.local
S15nfs-common S19atd S19sshS22rmnologin
S17rsyslog S19cron S20exim4 S22stop-bootlogd
In this case, After booting finished, I was able to find port 80 in tcp6 lines
of netstat -an.
I didn't need # /etc/init.d/apache2 restart.
But a few seconds elapse at the stage of exim4.
Therefore,
..Case 2..
/etc/rc.local is same as Case 1.
And I executed,
# mv /etc/rc2.d/{S20,K80}exim4
# update-rc.d exim4 defaults
# ls /etc/rc2.d/
K01exim4 S17rsyslog S19cron S21bootlogs
README S18apache2 S19openbsd-inetd S22rc.local
S14portmap S19acpidS19spamassassin S22rmnologin
S15nfs-common S19atd S19sshS22stop-bootlogd
In this case, After booting finished, I was unable to find port 80 in tcp6
lines.
I needed # /etc/init.d/apache2 restart.
All that you really need for ipv6 is the ::1 ... line. Having one
line less on your X-less box won't make a difference.
Is this your entire /etc/hosts? Don;t you have any ipv4 settings?!
The entire text is the following:
127.0.0.1 localhost
192.168.xx.xx .xxx.xx
# The following lines are desirable for IPv6 capable hosts
::1 ip6-localhost ip6-loopback
fe00::0 ip6-localnet
ff00::0 ip6-mcastprefix
ff02::1 ip6-allnodes
ff02::2 ip6-allrouters
Thanks,
Satoru
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20121007040120.3f9b3d6b.trn...@otip.jp
Re: apache2's handling of IP version 6
On Thu, Oct 04, 2012 at 04:53:23AM +0900, Satoru Otsubo wrote: Syntax error on line 2 of /etc/apache2/ports.conf What about this syntax error? -- If you're not careful, the newspapers will have you hating the people who are being oppressed, and loving the people who are doing the oppressing. --- Malcolm X -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121005100437.GB26889@tal
Re: apache2's handling of IP version 6
On 10/4/2012 12:06 PM, Tom H wrote: On Thu, Oct 4, 2012 at 7:23 AM, Stan Hoeppner s...@hardwarefreak.com wrote: On 10/4/2012 3:46 AM, Rick Thomas wrote: On Oct 3, 2012, at 8:40 PM, Satoru Otsubo wrote: But the phenomena are same, that is, When booting my PC, apache2 failed to start. And when I executed the following: # /etc/init.d/apache2 restart apache2 started successfully with the dual stack. Why this phenomena happens ? Is the apache2 daemon starting before the ipv6 part of the network configuration is completely up? His log error does seem to indicate Apache is starting before the ipv6 stack is available--one of the downsides of parallel init. Where did you see the log? The OP pasted this in his 2nd post: Starting web server: apache2[Thu Oct 04 01:02:03 2012]: [crit] (EAI 9)Address family for hostname not supported: Alloc_listener: failed to set up sockaddr for ::: Syntax error on line 2 of /etc/apache2/ports.conf Listen setup failed. Action 'start' failed. The Apache error log may have more information. failed! I think Chris Bannister is onto something. I previously overlooked the syntax error msg. A syntax error would explain the socket setup failure. Probably something as simple as a typo in ports.conf. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/506ebc4f.3060...@hardwarefreak.com
Re: apache2's handling of IP version 6
Hi, I'm initiator of this thread. On Thu, Oct 04, 2012 at 04:53:23AM +0900, Satoru Otsubo wrote: Syntax error on line 2 of /etc/apache2/ports.conf What about this syntax error? Syntax error on line 2 is the following: Listen [:::]:80 By the way, I thought the following concerning the problem in my CGI PC (my server PC): Internet-Router--Lan-my PC When booting PC, apache2 decides that ipv6 is Ok if it gets the ipv6's information and ipv6 is not Ok if it can't get the ipv6's information. And apache2 gets the ipv6's information from the router's RA(Router Advertisements). But the router can't send RA immediately. Therefore when apache2 decides on whether ipv6 is Ok, it can't get the information yet. Therefore it decides that ipv6 is not Ok. Therefore in case of dual stack, apache2 issues an error because ports.conf is as follows: Listen a.b.c.d:80 Listen [:::]:80 Listen a.b.c.d:443 Listen [:::]:443 (Because ipv6 address explicitely exists, apache2 issues an error when it can't get ipv6's information.) In case of ports.conf being the following: Listen 80 Listen 443, apache2 falls back to listen only ipv4 ports because no address format exists. In fact, in boot sequence, I scroll locked by clicking the Scroll Lock key just before apache2's line appears, and after some moments, I again click the Scroll Lock to restart the boot sequence. Then after booting completed, I executed, # netstat -an And I found there exists Listen 80 and Listen 443 in the tcp6 lines. My thought is wrong ? By the way, on the PC of Squeeze desktop install (my second PC), After booting completed, on command of # netstat -an , I can always find Listen 80 and Listen 443 in the tcp6 lines. (Without # /etc/init.d/apache2 restart) Why in the desktop environment, # /etc/init.d/apache2 restart is not needed ? Is apache2 of the desktop environment executed in the other way than CGI environment ? In fact, I found, /etc/hosts of my GUI PC includes # The following lines are desirable for IPv6 capable hosts ::1 localhost ip6-localhost ip6-loopback fe00::0 ip6-localnet ff00::0 ip6-mcastprefix ff02::1 ip6-allnodes ff02::2 ip6-allrouters ff02::3 ip6-allhosts whereas, /etc/hosts of my CUI PC only includes # The following lines are desirable for IPv6 capable hosts ::1 ip6-localhost ip6-loopback fe00::0 ip6-localnet ff00::0 ip6-mcastprefix ff02::1 ip6-allnodes ff02::2 ip6-allrouters Thanks, Satoru -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121006014214.0567fd4e.trn...@otip.jp
Re: apache2's handling of IP version 6
On Oct 3, 2012, at 8:40 PM, Satoru Otsubo wrote: But the phenomena are same, that is, When booting my PC, apache2 failed to start. And when I executed the following: # /etc/init.d/apache2 restart apache2 started successfully with the dual stack. Why this phenomena happens ? Is the apache2 daemon starting before the ipv6 part of the network configuration is completely up? Try putting ipv6 in your /etc/modules file... Rick -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/edf66603-9c79-4cfd-bd22-8379bbe5e...@pobox.com
Re: apache2's handling of IP version 6
On 10/4/2012 3:46 AM, Rick Thomas wrote: On Oct 3, 2012, at 8:40 PM, Satoru Otsubo wrote: But the phenomena are same, that is, When booting my PC, apache2 failed to start. And when I executed the following: # /etc/init.d/apache2 restart apache2 started successfully with the dual stack. Why this phenomena happens ? Is the apache2 daemon starting before the ipv6 part of the network configuration is completely up? His log error does seem to indicate Apache is starting before the ipv6 stack is available--one of the downsides of parallel init. Try putting ipv6 in your /etc/modules file... -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/506d7194.4080...@hardwarefreak.com
Re: apache2's handling of IP version 6
Thank you, Rick Is the apache2 daemon starting before the ipv6 part of the network configuration is completely up? Try putting ipv6 in your /etc/modules file... My PC: Linux squeeze Japanese version clean install. Current kernel version: 2.6.32-5-686 After I executed, # /etc/init.d/apache2 restart, I tried, # lsmod | grep -i ipv6 But nothing is displayed. kernel version 2.6.32-5-686 does, from the begining, have ipv6 as an integrated part, not as a module ? If so, is there a possibility that the apache2 daemon is starting before the ipv6 part of the network configuration is completely up ? Satoru -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121004222950.a254d00b.trn...@otip.jp
RE: apache2's handling of IP version 6
Hello Satoru, Is the apache2 daemon starting before the ipv6 part of the network configuration is completely up? Try putting ipv6 in your /etc/modules file... My PC: Linux squeeze Japanese version clean install. Current kernel version: 2.6.32-5-686 After I executed, # /etc/init.d/apache2 restart, I tried, # lsmod | grep -i ipv6 But nothing is displayed. kernel version 2.6.32-5-686 does, from the begining, have ipv6 as an integrated part, not as a module ? If so, is there a possibility that the apache2 daemon is starting before the ipv6 part of the network configuration is completely up ? As of Debian Squeeze the ipv6 support is integrated at kernel level and no longer a module. Just do a ifconfig eth0 or ip addr show eth0 and you will always see an ipv6 line. Putting ipv6 in \etc\modprobe.d\blacklist.conf to prevent ipv6 support also no longer works as some may have noticed. Bonno Bloksma -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/89d1798a7351d040b4e74e0a043c69d70f076...@einexch-01.tio.nl
Re: apache2's handling of IP version 6
On Thu, Oct 4, 2012 at 7:23 AM, Stan Hoeppner s...@hardwarefreak.com wrote: On 10/4/2012 3:46 AM, Rick Thomas wrote: On Oct 3, 2012, at 8:40 PM, Satoru Otsubo wrote: But the phenomena are same, that is, When booting my PC, apache2 failed to start. And when I executed the following: # /etc/init.d/apache2 restart apache2 started successfully with the dual stack. Why this phenomena happens ? Is the apache2 daemon starting before the ipv6 part of the network configuration is completely up? His log error does seem to indicate Apache is starting before the ipv6 stack is available--one of the downsides of parallel init. Where did you see the log? apache2 depends on $network. Wouldn't the ipv6 stack be available after the network's brought up? Doesn't Required-Start: $local_fs $remote_fs $network $syslog $named mean that apache2 starts once $network is satisfied/reached (meaning, I hope!, that the network's up). -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CAOdo=swdogc8qib7acatrnsb5ucoor6wmto9z_chipkkw8x...@mail.gmail.com
Re: apache2's handling of IP version 6
On 10/3/2012 9:08 AM, Satoru Otsubo wrote: I'm Satoru Otsubo I have some problems concerning apache2's handling of IP version 6. My PC: Linux squeeze Japanese version clean install. I use the normal apache2 package prepared by debian maintainers, that is, I did not compile apache2. This is server use, that is, no Grafical User Interface. (Current kernel version: 2.6.32-5-686) I think apache2 is using IPv4-mapped addresses, that is, apache2 lets only IPv6's ports open, and apache2 handles both the IPv6 and IPv4 tcp packets from IPv6's ports. Therefore I think apache2 normally has to listen IPv6's ports after booting PC. But, (i) After apache2 starts by booting my PC, apache2 listens only IPv4's 80 and 443 tcp ports. (ii) Then after I execute # /etc/init.d/apache2 restart, apache2 becomes normal, that is, apache2 becomes to listen only IPv6's 80 and 443 tcp ports. I have repeated many times. But after booting, apache2 always listens only IPv4's 80 and 443 tcp ports. And after # /etc/init.d/apache2 restart, apache2 always becomes normal. Why this phenoma happens ? In what situations apache2 falls back to use IPv4's 80 and 443 tcp ports instead of IPv6's 80 and 443 tcp ports ? If I know these situations, I might let apache2 listen IPv6's 80 and 443 tcp ports after booting my PC, by repairing files, etc. of my PC. Read this: http://www.cyberciti.biz/faq/ipv6-apache-configuration-tutorial/ -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/506c4e94.1020...@hardwarefreak.com
Re: apache2's handling of IP version 6
Thanks, Stan Read this: http://www.cyberciti.biz/faq/ipv6-apache-configuration-tutorial/ I modified /etc/apache2/ports.conf as follows: Listen a.b.c.d:80 Listen [:::]:80 Listen a.b.c.d:443 Listen [:::]:443 (where, a.b.c.d = my web server's IPv4 address ::: = my web server's IPv6 address) And I rebooted my PC. Then I found the following in the boot sequence: Starting web server: apache2[Thu Oct 04 01:02:03 2012]: [crit] (EAI 9)Address family for hostname not supported: Alloc_listener: failed to set up sockaddr for ::: Syntax error on line 2 of /etc/apache2/ports.conf Listen setup failed. Action 'start' failed. The Apache error log may have more information. failed! (There was nothing related to this error in /var/log/apache2/error.log) Then I executed the following: # /etc/init.d/apache2 restart And the Apache started successfully with the dual stack: Why apache2 succeeded to start when I executed # /etc/init.d/apache2 restart, after apache2 failed to start when booting ? Is this because there are any wrong in my apache2's settings ? Satoru -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121004045323.c81902fa.trn...@otip.jp
Re: apache2's handling of IP version 6
On Wed, Oct 3, 2012 at 3:53 PM, Satoru Otsubo trn...@otip.jp wrote: Thanks, Stan Read this: http://www.cyberciti.biz/faq/ipv6-apache-configuration-tutorial/ I modified /etc/apache2/ports.conf as follows: Listen a.b.c.d:80 Listen [:::]:80 Listen a.b.c.d:443 Listen [:::]:443 (where, a.b.c.d = my web server's IPv4 address ::: = my web server's IPv6 address) And I rebooted my PC. Then I found the following in the boot sequence: Starting web server: apache2[Thu Oct 04 01:02:03 2012]: [crit] (EAI 9)Address family for hostname not supported: Alloc_listener: failed to set up sockaddr for ::: Syntax error on line 2 of /etc/apache2/ports.conf Listen setup failed. Action 'start' failed. The Apache error log may have more information. failed! (There was nothing related to this error in /var/log/apache2/error.log) Then I executed the following: # /etc/init.d/apache2 restart And the Apache started successfully with the dual stack: Why apache2 succeeded to start when I executed # /etc/init.d/apache2 restart, after apache2 failed to start when booting ? Did you see, on the page that Stan sent you that you need a VirtualHost stanza for ipv6: VirtualHost 74.86.48.99 VirtualHost [2607:f0d0:1002:11::4] or VirtualHost 74.86.48.99 [2607:f0d0:1002:11::4] (Also, Listen 80 covers both ipv4 and ipv6.) -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CAOdo=sx97kykrtocvex_q8tw0jwwnybsjc3ump67gyxh+ev...@mail.gmail.com
Re: apache2's handling of IP version 6
Thanks, Stan Did you see, on the page that Stan sent you that you need a VirtualHost stanza for ipv6: VirtualHost 74.86.48.99 VirtualHost [2607:f0d0:1002:11::4] or VirtualHost 74.86.48.99 [2607:f0d0:1002:11::4] (Also, Listen 80 covers both ipv4 and ipv6.) I changed ports.conf as described on the previous mailing. And I changed files in /etc/apache2/sites-available/ as follows: Concerning *80, as follows: NameVirtualHost a.b.c.d:80 NameVirtualHost [:::]:80 VirtualHost a.b.c.d:80 [:::]:80 And Concerning *443, as follows: NameVirtualHost a.b.c.d:443 NameVirtualHost [:::]:443 VirtualHost a.b.c.d:443 [:::]:443 But the phenomena are same, that is, When booting my PC, apache2 failed to start. And when I executed the following: # /etc/init.d/apache2 restart apache2 started successfully with the dual stack. Why this phenomena happens ? Satoru -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121004124059.0495fc78.trn...@otip.jp
