DOvecot requires both IPv4 and IPV6 to start
I think we are off topic.
We have a network policy that until such time as we have fully implemented the same security we have on our IPV4 network we simply do not allow IPV6 as it is another attack vector we are not fully upto speed on. So we block all IPV6 at the perimeter and on all switches, firewalls and routers. So if you connect to our network with a device no IP6 is configured and only IPV4 is allowed. We will get there!
But that is not this issue. The issue is that dovecot is assuming that IPV6 is there and crashes out if it's not. Hacking the config to only listen on IP4 solves the problem but not while installing from standard repos to install the currently supported version.
Cheers Spart
On 04/09/2023 11:12, TWHG Technical via dovecot wrote:
But that is not this issue. The issue is that dovecot is assuming that IPV6 is there and crashes out if it's not. Hacking the config to only listen on IP4 solves the problem but not while installing from standard repos to install the currently supported version.
dovecot can handle an IPv4-only setup, you just have to tell it to (by setting a custom 'listen' config entry, as you have already found out).
Your setup is non-standard, so it's expected that you'll have to make changes to accommodate that. Expecting that dovecot changes its default because of your particularities may be asking a little too much.
PS: It would be easier to follow the discussion if you actually replied to the messages (quoting the relevant parts) instead of sending a new message. But be sure to use a client that sets in-reply-to: or references: headers so that the thread is not broken.
-- Life is too short to stuff a mushroom. -- Storm Jameson
Eduardo M KALINOWSKI eduardo@kalinowski.com.br
On 2023-09-04 08:58, Eduardo M KALINOWSKI via dovecot wrote:
On 04/09/2023 11:12, TWHG Technical via dovecot wrote:
But that is not this issue. The issue is that dovecot is assuming that IPV6 is there and crashes out if it's not. Hacking the config to only listen on IP4 solves the problem but not while installing from standard repos to install the currently supported version.
dovecot can handle an IPv4-only setup, you just have to tell it to (by setting a custom 'listen' config entry, as you have already found out).
Your setup is non-standard, so it's expected that you'll have to make changes to accommodate that. Expecting that dovecot changes its default because of your particularities may be asking a little too much.
PS: It would be easier to follow the discussion if you actually replied to the messages (quoting the relevant parts) instead of sending a new message. But be sure to use a client that sets in-reply-to: or references: headers so that the thread is not broken.
However, I 'get' this persons' opinion, from a developers perspective.
The system should either run, or provide a clear reason why it didn't startup (that reason could be .. You have selected * but IPv6 is not available). Doesn't really matter what the dependency is, whether a missing package, or a service not responding, there should be sane checks, and turning off IPv6 is probably a lot more popular than you think, given the increased attack vector and other observed issues.
But of course, the listen directive can easily be modified. Just harder for newbies looking for an 'out of the box' solution.
-- "Catch the Magic of Linux..."
Michael Peddemors, President/CEO LinuxMagic Inc. Visit us at http://www.linuxmagic.com @linuxmagic A Wizard IT Company - For More Info http://www.wizard.ca "LinuxMagic" a Registered TradeMark of Wizard Tower TechnoServices Ltd.
604-682-0300 Beautiful British Columbia, Canada
This email and any electronic data contained are confidential and intended solely for the use of the individual or entity to which they are addressed. Please note that any views or opinions presented in this email are solely those of the author and are not intended to represent those of the company.
I will reply to the email rather than press reply on the web mailing list page. Probably got this wrong too!
I don't consider our setup non-standard. It is a standard IPV4 setup. We will support IPV6 once our network security is reworked and everyone has a full understanding of IPV6 and it's challenges. An IPV6 stack on the box is an additional attack vector we are not setup up for at present. We will get there.
Simply assuming IPV6 is there and then crashing out when it's not is not useful when installing as part of another package or from the supported OS repos. Yes, I understand we can hack the config to workaround this. But only after the install has failed.
An example would be installing a virtualmin server for instance. A LAMP stack and all Packages are installed from the repos as part of an installer script. Dovecot crashes out as no IPV6 and that leaves the server in an unusable state.
I believe it really would be good practice to fallback to IPV4 only listeners if IPV6 listener fails. Rather than crash out.
Thank you for your consideration. Cheers
----- Original Message ----- From: "Michael Peddemors" michael@linuxmagic.com To: "dovecot" dovecot@dovecot.org Sent: Monday, 4 September, 2023 19:36:45 Subject: Re: DOvecot requires both IPv4 and IPV6 to start
On 2023-09-04 08:58, Eduardo M KALINOWSKI via dovecot wrote:
On 04/09/2023 11:12, TWHG Technical via dovecot wrote:
But that is not this issue. The issue is that dovecot is assuming that IPV6 is there and crashes out if it's not. Hacking the config to only listen on IP4 solves the problem but not while installing from standard repos to install the currently supported version.
dovecot can handle an IPv4-only setup, you just have to tell it to (by setting a custom 'listen' config entry, as you have already found out).
Your setup is non-standard, so it's expected that you'll have to make changes to accommodate that. Expecting that dovecot changes its default because of your particularities may be asking a little too much.
PS: It would be easier to follow the discussion if you actually replied to the messages (quoting the relevant parts) instead of sending a new message. But be sure to use a client that sets in-reply-to: or references: headers so that the thread is not broken.
However, I 'get' this persons' opinion, from a developers perspective.
The system should either run, or provide a clear reason why it didn't startup (that reason could be .. You have selected * but IPv6 is not available). Doesn't really matter what the dependency is, whether a missing package, or a service not responding, there should be sane checks, and turning off IPv6 is probably a lot more popular than you think, given the increased attack vector and other observed issues.
But of course, the listen directive can easily be modified. Just harder for newbies looking for an 'out of the box' solution.
-- "Catch the Magic of Linux..."
Michael Peddemors, President/CEO LinuxMagic Inc. Visit us at http://www.linuxmagic.com @linuxmagic A Wizard IT Company - For More Info http://www.wizard.ca "LinuxMagic" a Registered TradeMark of Wizard Tower TechnoServices Ltd.
604-682-0300 Beautiful British Columbia, Canada
This email and any electronic data contained are confidential and intended solely for the use of the individual or entity to which they are addressed. Please note that any views or opinions presented in this email are solely those of the author and are not intended to represent those of the company.
dovecot mailing list -- dovecot@dovecot.org To unsubscribe send an email to dovecot-leave@dovecot.org
On 9/4/23 13:42, TWHG Technical via dovecot wrote:
I believe it really would be good practice to fallback to IPV4 only listeners if IPV6 listener fails. Rather than crash out.
On my previous reply, I had not yet read the other messages. They were part of another thread in Thunderbird that I did not see at first.
After reading that and doing a little investigation, I can see that the comment in my main dovecot config is correct ... * means all Ipv4 interfaces. The default config uses '*, ::' for the listen option.
Dovecot does at least clearly log the problem when IPv6 is configured and no IPv6 is available:
Sep 4 14:26:41 bilbo dovecot: master: Error: socket() failed: Address family not supported by protocol Sep 4 14:26:41 bilbo dovecot: master: Error: service(managesieve-login): listen(::, 4190) failed: Address family not supported by protocol Sep 4 14:26:41 bilbo dovecot: master: Error: socket() failed: Address family not supported by protocol Sep 4 14:26:41 bilbo dovecot: master: Error: service(pop3-login): listen(::, 110) failed: Address family not supported by protocol Sep 4 14:26:41 bilbo dovecot: master: Error: socket() failed: Address family not supported by protocol Sep 4 14:26:41 bilbo dovecot: master: Error: service(pop3-login): listen(::, 995) failed: Address family not supported by protocol Sep 4 14:26:41 bilbo dovecot: master: Error: socket() failed: Address family not supported by protocol Sep 4 14:26:41 bilbo dovecot: master: Error: service(imap-login): listen(::, 143) failed: Address family not supported by protocol Sep 4 14:26:41 bilbo dovecot: master: Error: socket() failed: Address family not supported by protocol Sep 4 14:26:41 bilbo dovecot: master: Error: service(imap-login): listen(::, 993) failed: Address family not supported by protocol Sep 4 14:26:41 bilbo dovecot: master: Fatal: Failed to start listeners
The default config does explicitly ask for ipv6, so Dovecot did try (and fail) to start with the requested config, and most importantly, it did report the problem.
I am not sure that it should have failed to start, though. I think that ipv4/ipv6 might need to be a special case, where a program logs an error, but continues to start up and run with support for the other address family.
I disable ipv6 on all my servers by adding "ipv6.disable=1" to the kernel commandline in /etc/default/grub. I do it for two reasons. 1) Some software can behave strangely when ipv6 is enabled. 2) I am not using ipv6 and do not want my programs doing something unexpected because an IPv6 interface was contacted. #1 is probably very rare in recent years, but it was common in the past. I know that a lot of other admins routinely disable ipv6 on their machines for similar reasons.
An interesting read:
https://cr.yp.to/djbdns/ipv6mess.html
Thanks, Shawn
Citeren Shawn Heisey elyograg@elyograg.org:
An interesting read:
FUD.
This document was written something like two decades ago and pretty
much all arguments against IPv6 are no longer relevant anymore. In
most cases nowadays, the user experience for IPv6 will be far superior
over CGNAT connections which unfortunately become the standard due to
the increasing cost of IPv4 addresses. Running services over IPv6 may
have been experimental and prone with not so fun problems to diagnose
in 2013, but today this is no longer the case.
On 9/4/23 15:40, Arjen de Korte wrote:
Citeren Shawn Heisey elyograg@elyograg.org:
An interesting read:
FUD.
This document was written something like two decades ago and pretty much all arguments against IPv6 are no longer relevant anymore. In most cases nowadays, the user experience for IPv6 will be far superior over CGNAT connections which unfortunately become the standard due to the increasing cost of IPv4 addresses. Running services over IPv6 may have been experimental and prone with not so fun problems to diagnose in 2013, but today this is no longer the case.
So you're saying that I can change my entire home network to IPv6, eliminate IPv4 entirely, and I will have no problem connecting to sites like bbc.com or cnn.com, which have no AAAA record? My ISP is Comcast, which does support IPv6, though I have it entirely disabled.
If there is a way for an IPv6 computer to connect to websites with no IPv6 addresses, how could it possibly happen transparently?
Thanks, Shawn
. On 05/09/2023 02:39, Shawn Heisey wrote:
[...] So you're saying that I can change my entire home network to IPv6, eliminate IPv4 entirely, and I will have no problem connecting to sites
On 9/4/23 15:40, Arjen de Korte wrote: like bbc.com or cnn.com, which have no AAAA record? My ISP is Comcast, which does support IPv6, though I have it entirely disabled.
Yes. It's possible.
But cnn.com and bbc.com are not a good example because they are behing IPv6 nodes (Thanks to some CDNs)
Check http://paste.debian.net/1290953/ .
I used github.com as a current example though and you can see it works in my network.
If there is a way for an IPv6 computer to connect to websites with no IPv6 addresses, how could it possibly happen transparently? Transparently?No. But at least you have several transition mechanisms such as NAT64+DNS64 . I used both in some networks like the one in my example.
And yes, I have some devices running IPv6 only and I have IPv4 at the edge only of that network.
-- Willy Manga @ongolaboy https://ongola.blogspot.com/
On 9/4/23 12:36, Michael Peddemors wrote:
However, I 'get' this persons' opinion, from a developers perspective.
The system should either run, or provide a clear reason why it didn't startup (that reason could be .. You have selected * but IPv6 is not available). Doesn't really matter what the dependency is, whether a missing package, or a service not responding, there should be sane checks, and turning off IPv6 is probably a lot more popular than you think, given the increased attack vector and other observed issues.
But of course, the listen directive can easily be modified. Just harder for newbies looking for an 'out of the box' solution.
The following is in my /etc/dovecot/dovecot.conf file. Ubuntu 22, using the official dovecot APT repo:
# A comma separated list of IPs or hosts where to listen in for connections. # "*" listens in all IPv4 interfaces, "::" listens in all IPv6 interfaces. # If you want to specify non-default ports or anything more complex, # edit conf.d/master.conf. listen = *
My dovecot (version 2:2.3.20-3+ubuntu20.04) works just fine. I have IPv6 completely disabled on the server.
The comment in my config file indicates that * should be all available IPv4 interfaces, which contradicts what you said above. Maybe on my version the comment is right and for newer versions it isn't?
I have found a bunch of software that will refuse to run if IPv6 is disabled unless configured explicitly to use v4 and disable v6. I find this trend disturbing. In one case (atftpd) it took me a very long time to determine that the lack of an IPv6 interface was the cause for the program not starting. Once I figured that out, I just added "--ipv4" to the /etc/default/atftpd file and it started working.
Even if * really means "all interfaces" and not "all IPv4 interfaces" as the comment indicates ... if no IPv6 interfaces are found, dovecot should proceed with the interfaces it finds, not fail when none of those interfaces have IPv6 addressing.
To the OP: Was it obvious in logs that IPv6 was the problem? In a lot of cases programs that refuse to start without IPv6 being available will do so silently. It's the silently part that's the real problem with this trend.
Thanks, Shawn
Shawn,
That config is not the standard as from the repos. As the standard one has that line commented out signifying it is the default and the default is now:
listen = *, ::
The :: is the issue here as if there is no IPV6 stack and dovecot crashes as, as a default it expects IPV6 to be there or else!
Yes, it was obvious from the systemctl status dovecot output.
I include for your amusement.
Processing triggers for dovecot-core (1:2.3.16+dfsg1-3ubuntu2.2) ... Job for dovecot.service failed because the control process exited with error code. See "systemctl status dovecot.service" and "journalctl -xeu dovecot.service" for details. invoke-rc.d: initscript dovecot, action "restart" failed. ^[[0;1;31m×^[[0m dovecot.service - Dovecot IMAP/POP3 email server Loaded: loaded (^[]8;;file://newcloud.twhg.co.uk/lib/systemd/system/dovecot.service^G/lib/systemd/system/dovecot.service^[]8;;^G; enabled;> Active: ^[[0;1;31mfailed^[[0m (Result: exit-code) since Mon 2023-09-04 10:45:29 UTC; 9ms ago Docs: ^[]8;;man:dovecot(1)^Gman:dovecot(1)^[]8;;^G ^[]8;;https://doc.dovecot.org/^Ghttps://doc.dovecot.org/^[]8;;^G Process: 63336 ExecStart=/usr/sbin/dovecot -F ^[[0;1;31m(code=exited, status=89)^[[0m Main PID: 63336 (code=exited, status=89) CPU: 30ms
Sep 04 10:45:29 ourdomain dovecot[63336]: ^[[0;1;31m^[[0;1;39m^[[0;1;31mmaster: Error: service(imap-login): listen(::, 143) failed: A> Sep 04 10:45:29 ourdomain dovecot[63336]: Error: socket() failed: Address family not supported by protocol Sep 04 10:45:29 ourdomain dovecot[63336]: Error: service(imap-login): listen(::, 993) failed: Address family not supported by protocol Sep 04 10:45:29 ourdomain dovecot[63336]: ^[[0;1;31m^[[0;1;39m^[[0;1;31mmaster: Error: socket() failed: Address family not supported > Sep 04 10:45:29 ourdomain dovecot[63336]: ^[[0;1;31m^[[0;1;39m^[[0;1;31mmaster: Error: service(imap-login): listen(::, 993) failed: A> Sep 04 10:45:29 ourdomain dovecot[63336]: Fatal: Failed to start listeners Sep 04 10:45:29 ourdomain dovecot[63336]: ^[[0;1;31m^[[0;1;39m^[[0;1;31mmaster: Fatal: Failed to start listeners^[[0m Sep 04 10:45:29 ourdomain systemd[1]: ^[[0;1;39m^[[0;1;31m^[[0;1;39mdovecot.service: Main process exited, code=exited, status=89/n/a^> Sep 04 10:45:29 ourdomain systemd[1]: ^[[0;1;38;5;185m^[[0;1;39m^[[0;1;38;5;185mdovecot.service: Failed with result 'exit-code'.^[[0m Sep 04 10:45:29 ourdomain systemd[1]: ^[[0;1;31m^[[0;1;39m^[[0;1;31mFailed to start Dovecot IMAP/POP3 email server.^[[0m dpkg: error processing package dovecot-core (--configure):
Cheers
----- Original Message ----- From: "Shawn Heisey" elyograg@elyograg.org To: "dovecot" dovecot@dovecot.org Sent: Monday, 4 September, 2023 21:13:42 Subject: Re: DOvecot requires both IPv4 and IPV6 to start
On 9/4/23 12:36, Michael Peddemors wrote:
However, I 'get' this persons' opinion, from a developers perspective.
The system should either run, or provide a clear reason why it didn't startup (that reason could be .. You have selected * but IPv6 is not available). Doesn't really matter what the dependency is, whether a missing package, or a service not responding, there should be sane checks, and turning off IPv6 is probably a lot more popular than you think, given the increased attack vector and other observed issues.
But of course, the listen directive can easily be modified. Just harder for newbies looking for an 'out of the box' solution.
The following is in my /etc/dovecot/dovecot.conf file. Ubuntu 22, using the official dovecot APT repo:
# A comma separated list of IPs or hosts where to listen in for connections. # "*" listens in all IPv4 interfaces, "::" listens in all IPv6 interfaces. # If you want to specify non-default ports or anything more complex, # edit conf.d/master.conf. listen = *
My dovecot (version 2:2.3.20-3+ubuntu20.04) works just fine. I have IPv6 completely disabled on the server.
The comment in my config file indicates that * should be all available IPv4 interfaces, which contradicts what you said above. Maybe on my version the comment is right and for newer versions it isn't?
I have found a bunch of software that will refuse to run if IPv6 is disabled unless configured explicitly to use v4 and disable v6. I find this trend disturbing. In one case (atftpd) it took me a very long time to determine that the lack of an IPv6 interface was the cause for the program not starting. Once I figured that out, I just added "--ipv4" to the /etc/default/atftpd file and it started working.
Even if * really means "all interfaces" and not "all IPv4 interfaces" as the comment indicates ... if no IPv6 interfaces are found, dovecot should proceed with the interfaces it finds, not fail when none of those interfaces have IPv6 addressing.
To the OP: Was it obvious in logs that IPv6 was the problem? In a lot of cases programs that refuse to start without IPv6 being available will do so silently. It's the silently part that's the real problem with this trend.
Thanks, Shawn
dovecot mailing list -- dovecot@dovecot.org To unsubscribe send an email to dovecot-leave@dovecot.org
On 9/4/23 1:41 PM, TWHG Technical via dovecot wrote:
Sep 04 10:45:29 ourdomain dovecot[63336]: ^[[0;1;31m^[[0;1;39m^[[0;1;31mmaster: Error: service(imap-login): listen(::, 143) failed: A> Sep 04 10:45:29 ourdomain dovecot[63336]: Error: socket() failed: Address family not supported by protocol Sep 04 10:45:29 ourdomain dovecot[63336]: Error: service(imap-login): listen(::, 993) failed: Address family not supported by protocol Sep 04 10:45:29 ourdomain dovecot[63336]: ^[[0;1;31m^[[0;1;39m^[[0;1;31mmaster: Error: socket() failed: Address family not supported > Sep 04 10:45:29 ourdomain dovecot[63336]: ^[[0;1;31m^[[0;1;39m^[[0;1;31mmaster: Error: service(imap-login): listen(::, 993) failed: A> Sep 04 10:45:29 ourdomain dovecot[63336]: Fatal: Failed to start listeners Sep 04 10:45:29 ourdomain dovecot[63336]: ^[[0;1;31m^[[0;1;39m^[[0;1;31mmaster: Fatal: Failed to start listeners^[[0m Sep 04 10:45:29 ourdomain systemd[1]: ^[[0;1;39m^[[0;1;31m^[[0;1;39mdovecot.service: Main process exited, code=exited, status=89/n/a^> Sep 04 10:45:29 ourdomain systemd[1]: ^[[0;1;38;5;185m^[[0;1;39m^[[0;1;38;5;185mdovecot.service: Failed with result 'exit-code'.^[[0m Sep 04 10:45:29 ourdomain systemd[1]: ^[[0;1;31m^[[0;1;39m^[[0;1;31mFailed to start Dovecot IMAP/POP3 email server.^[[0m dpkg: error processing package dovecot-core (--configure):
I can understand why you dislike the cascade of failure messages. I dislike them as well. Of course, I've been following the RFCs for longer than I can remember, as well as doing network programming in the C language, among others, so I recognize "Address family not supported" for the error message that it is.
Here's a pro tip: your concern is about security...so use security tools to implement what you want. For Ubuntu, that's the ufw(8) firewall system. (Other distributions use other utilities, like iptables(8), but that's beside the point.)
Want to lock out IPv6? Do it with ufw firewall settings. That would cause the least amount of breakage. More stuff will work out of the box using the firewall to block ne'er-do-wells.
I'll also speak as a long-time software developer: you HAVE to make assumptions about the underlying system in order to stay sane, let alone concentrate on making the software perform its function correctly. In the Unix environment, particularly in shell scripts, you depend on certain binaries being available. It's bad enough trying to design software to run in different distributions or even operating environments (Linux v Unix v Mac v Windows) without having to second-guess administrators who remove tools willy-nilly.
The above should be part of System Administration 101. Which I never took, but had to learn in the school of hard knocks. And I do know that the certification bodies don't test for that particular nugget. Which is why we have mailing lists to help people with sore toes.
On 9/4/23 12:20, Eduardo M KALINOWSKI via dovecot wrote:
On 04/09/2023 11:12, TWHG Technical via dovecot wrote:
But that is not this issue. The issue is that dovecot is assuming that IPV6 is there and crashes out if it's not. Hacking the config to only listen on IP4 solves the problem but not while installing from standard repos to install the currently supported version.
dovecot can handle an IPv4-only setup, you just have to tell it to (by setting a custom 'listen' config entry, as you have already found out).
Your setup is non-standard, so it's expected that you'll have to make changes to accommodate that. Expecting that dovecot changes its default because of your particularities may be asking a little too much.
PS: It would be easier to follow the discussion if you actually replied to the messages (quoting the relevant parts) instead of sending a new message. But be sure to use a client that sets in-reply-to: or references: headers so that the thread is not broken.
Ah contraire, Eduardo. My location since 1984 is in the middle northern area of WV, USA, And I am a minimum of 150 kilometers from the nearest ipv6 enabled network connection. I'm not even sure my cable modem, furnished by Shentel about 7 to 10 years ago, even can be configured to handle ipv6. Not my choice, except where I live, place is free & clear & has been for 24 years now, good neighbors in a small county seat town. Retired for 21 years, no reason to move unless I leave in a box.
By not accommodating the ipv6-less yet masses with a too bad, so sad attitude is unbecoming. You may have ipv6 at your router input, but there are millions not so lucky. You apparently have the power to fix it, please do so.
Thank you Eduardo.
Cheers, Gene Heskett.
"There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author, 1940) If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis Genes Web page http://geneslinuxbox.net:6309/
04.09.2023 23:23, gene heskett wrote: ..
Ah contraire, Eduardo. My location since 1984 is in the middle northern area of WV, USA, And I am a minimum of 150 kilometers from the nearest ipv6 enabled network connection. I'm not even sure my cable modem, furnished by Shentel about 7 to 10 years ago, even can be configured to handle ipv6.
Not my choice, except where I live, place is free & clear & has been for 24 years now, good neighbors in a small county seat town. Retired for 21 years, no reason to move unless I leave in a box.By not accommodating the ipv6-less yet masses with a too bad, so sad attitude is unbecoming. You may have ipv6 at your router input, but there are millions not so lucky. You apparently have the power to fix it, please do so.
This is apples and oranges. Lack of IPv6 conectivity might be quite common still, I dunno. But lack of IPv6 *support* in the system is very uncommon. For many years v6 worked to co-exists with v4 nicely, and if there's no v6 connectivity, to fall back to v4 transparently. It just works (but I must admit, this works less and less good, since fallback code paths are tested less and less often). If you disable something on your own system which is commonly used (v6 support), it is your task to deal with the consequences. Maintainers can help in some cases or can make this easier, but this is definitely not a priority, esp. once a trivial work-around exists (to configure a package to use v4-only).
/mjt
On 9/5/23 00:52, Michael Tokarev wrote:
04.09.2023 23:23, gene heskett wrote: ..
Ah contraire, Eduardo. My location since 1984 is in the middle northern area of WV, USA, And I am a minimum of 150 kilometers from the nearest ipv6 enabled network connection. I'm not even sure my cable modem, furnished by Shentel about 7 to 10 years ago, even can be configured to handle ipv6. Not my choice, except where I live, place is free & clear & has been for 24 years now, good neighbors in a small county seat town. Retired for 21 years, no reason to move unless I leave in a box.
By not accommodating the ipv6-less yet masses with a too bad, so sad attitude is unbecoming. You may have ipv6 at your router input, but there are millions not so lucky. You apparently have the power to fix it, please do so.
This is apples and oranges. Lack of IPv6 conectivity might be quite common still, I dunno. But lack of IPv6 *support* in the system is very uncommon. For many years v6 worked to co-exists with v4 nicely, and if there's no v6 connectivity, to fall back to v4 transparently.
I have yet to do an install since wheezy that did NOT require my manual intervention to make ipv4 work. So in that respect, the installers have failed miserably.
This Bookworm was by surprise, because a buster update did something to my user pw, locking me totally out, so I had no chance the read the installer notes, and the only recovery was taking my optical drive to one of the milling machines in the garage, going online and downloading the net-install iso and writing it to a dvd just to install bookworm on this machine. And several months later, no one can tell me why an app that needs access to my raid10 /home partition, takes from 30 secs up to 5 minutes after the mouse click, to open a file requestor accessing this raid10 /home. Why? What difference is there between the file access in buster, and the file access in bookworm? I've even forced a fsck on it using bookworms tools. No clues.
It just works (but I must admit, this works less and less
good, since fallback code paths are tested less and less often). If you disable something on your own system which is commonly used (v6 support), it is your task to deal with the consequences. Maintainers can help in some cases or can make this easier, but this is definitely not a priority, esp. once a trivial work-around exists (to configure a package to use v4-only).
/mjt
dovecot mailing list -- dovecot@dovecot.org To unsubscribe send an email to dovecot-leave@dovecot.org .
Cheers, Gene Heskett.
"There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author, 1940) If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis Genes Web page http://geneslinuxbox.net:6309/
participants (9)
-
Arjen de Korte
-
Eduardo M KALINOWSKI
-
gene heskett
-
Michael Peddemors
-
Michael Tokarev
-
Shawn Heisey
-
Stephen Satchell
-
TWHG Technical
-
Willy Manga