Language:
switch to room list switch to menu My folders
Go to page: First ... 49 50 51 52 [53] 54 55 56 57 ... Last
[#] Thu Jun 11 2015 05:28:41 EDT from dothebart @ Uncensored

Subject: Re: Suggestions for migrating Exchange/SBS 2011 to Citadel/Debian Jessie

[Reply] [ReplyQuoted] [Headers] [Print]

 

Wed Jun 10 2015 11:48:06 EDT from CarlGambolputty @ Uncensored Subject: Re: Suggestions for migrating Exchange/SBS 2011 to Citadel/Debian Jessie

Doi. For some reason, when I compiled my citadel packages I didn't get a citadel-client package. Thought it would be contained in one of the other ones. Herp derp. It's installed now, so that's out of the way.



Oh, it was moved out of the citserver package; I need to update the documentation.



[#] Thu Jun 11 2015 12:37:05 EDT from CarlGambolputty @ Uncensored

Subject: Re: Suggestions for migrating Exchange/SBS 2011 to Citadel/Debian Jessie

[Reply] [ReplyQuoted] [Headers] [Print]

@ dothebart: <meanie_face>

OK. I think I have it narrowed down some. I've been working on this issue for the past few days as time allowed.

The problem here is Webcit, or some dependency of it that it uses, failing to communicate with the AD, with the faliure on the end of the Citadel install, not the Debian server or the AD DC. As mentioned, I connected to the AD with Samba, and can pull tickets, get results from wbinfo, and can log onto the system with home folders using AD users. The only hiccup is that CTDBD, which is a file-sharing component of Samba, is unhappy living on my Jessie install, but that's due to my nodes/private IPs in /etc/ctdb/nodes not being declared (because I don't care), so it's not related AFAIK. 

I think the fault originates in citadel-server, since trying to use the Citadel client from the install box through the shell runs into the same issue; it pretends that users can't be seen. The fact that Webcit doesn't log me on is because the citadel server isn't feeding it information from the AD, because Citadel can't find it.

Backing this up are entries I'm finding in /var/log/syslog when I attempt to log in to the web portal; after webcit notes the log-in IP and matches the session with a cookie, it passes stuff over to the citserver process, which then barfs this up:

 

Jun 11 11:46:35 KRILLAN citserver[812]: [84][(not logged in)(0)] USER ise_adm

Jun 11 11:46:35 KRILLAN citserver[812]: LDAP bind DN: cn=queries,cn=users,dc=domain,dc=local

Jun 11 11:46:35 KRILLAN citserver[812]: LDAP: Cannot bind: Can't contact LDAP server (-1)

Jun 11 11:46:35 KRILLAN webcit[2492]: HTTP: 200 [0.001153] POST /ajax_login_username_password

 

(I'm not sure if citserver would write a different error string if it COULD find the server, but COULDN'T get a query return on the DN bind account.)
 
In addition, when I try running webcit as a direct command at the shell, rather than from /etc/init.d/webcit, I get a bunch of returns in console that look like this (where 192.168.1.2 represents the IP of our domain box):
 
nss_ldap: could not connect to any LDAP server as CN=ise admin,DC=integrated,DC=local - Can't contact LDAP server
nss_ldap: failed to bind to LDAP server ldapi://192.168.1.2/: Can't contact LDAP server
nss_ldap: reconnecting to LDAP server...
 
And so on and so forth. Note that this is the only time I get this message, in console or in logs. I am, by the way, using option '3' (non-standard LDAP) for authentication, though I tried it with option 2 (RFC-compliant LDAP) previously with the same result. I've also verified the bind account's creds on the Windows server; I even made an additional user (info above) with values matching the online guide for funsies, pointed to it with a fresh install, and had the same results.
 
The only other stand-out data point I've found is a WARNING in syslog when I restart the webcit service:
 
Jun 11 11:45:55 KRILLAN webcit[2492]: WARNING: getpwuid(65534): Success#012Group IDs will be incorrect.
Jun 11 11:45:55 KRILLAN webcit[2492]: Changing uid to 65534
Jun 11 11:45:55 KRILLAN webcit[2495]: Listening on socket 7
Jun 11 11:45:55 KRILLAN webcit[2495]: Requesting cipher list: DEFAULT
Jun 11 11:45:55 KRILLAN webcit[2495]: WARNING: getpwuid(65534): File exists#012Group IDs will be incorrect.
Jun 11 11:45:55 KRILLAN webcit[2495]: Changing uid to 65534
 
I've found a reference to this error on https://www.mail-archive.com/room_citadel_development@uncensored.citadel.org/msg09459.html, which I'm going to do when I have a moment, though I'm not sure if this is a relevant issue.
 
 
 
If anyone is interested in the syslog, let me know and I'll email it over (since this forums doesn't seem to support code boxes or post attachments). Again, I'd love suggestions on how to proceed since I'm at a bit of a brick wall, but past anything else I'd also be interested in suggestions on an optimal distro of linux to run this on, as on Saturday, if I'm still stuck I'm going to go nuclear option on this and start from scratch with something else than Debian Jessie.


[#] Thu Jun 11 2015 13:16:24 EDT from CarlGambolputty @ Uncensored

Subject: Re: Suggestions for migrating Exchange/SBS 2011 to Citadel/Debian Jessie

[Reply] [ReplyQuoted] [Headers] [Print]

@ dothebart: <meanie_face>

OK. I think I have it narrowed down some. I've been working on this issue for the past few days as time allowed.

The problem here is Webcit, or some dependency of it that it uses, failing to communicate with the AD, with the faliure on the end of the Citadel install, not the Debian server or the AD DC. As mentioned, I connected to the AD with Samba, and can pull tickets, get results from wbinfo, and can log onto the system with home folders using AD users. The only hiccup is that CTDBD, which is a file-sharing component of Samba, is unhappy living on my Jessie install, but that's due to my nodes/private IPs in /etc/ctdb/nodes not being declared (because I don't care), so it's not related AFAIK. 

I think the fault originates in citadel-server, since trying to use the Citadel client from the install box through the shell runs into the same issue; it pretends that users can't be seen. The fact that Webcit doesn't log me on is because the citadel server isn't feeding it information from the AD, because Citadel can't find it.

Backing this up are entries I'm finding in /var/log/syslog when I attempt to log in to the web portal; after webcit notes the log-in IP and matches the session with a cookie, it passes stuff over to the citserver process, which then barfs this up:

 

Jun 11 11:46:35 KRILLAN citserver[812]: [84][(not logged in)(0)] USER ise_adm

Jun 11 11:46:35 KRILLAN citserver[812]: LDAP bind DN: cn=queries,cn=users,dc=domain,dc=local

Jun 11 11:46:35 KRILLAN citserver[812]: LDAP: Cannot bind: Can't contact LDAP server (-1)

Jun 11 11:46:35 KRILLAN webcit[2492]: HTTP: 200 [0.001153] POST /ajax_login_username_password

 

(I'm not sure if citserver would write a different error string if it COULD find the server, but COULDN'T get a query return on the DN bind account.)
 
In addition, when I try running webcit as a direct command at the shell, rather than from /etc/init.d/webcit, I get a bunch of returns in console that look like this (where 192.168.1.2 represents the IP of our domain box):
 
nss_ldap: could not connect to any LDAP server as CN=ise admin,DC=integrated,DC=local - Can't contact LDAP server
nss_ldap: failed to bind to LDAP server ldapi://192.168.1.2/: Can't contact LDAP server
nss_ldap: reconnecting to LDAP server...
 
And so on and so forth. Note that this is the only time I get this message, in console or in logs. I am, by the way, using option '3' (non-standard LDAP) for authentication, though I tried it with option 2 (RFC-compliant LDAP) previously with the same result. I've also verified the bind account's creds on the Windows server; I even made an additional user (info above) with values matching the online guide for funsies, pointed to it with a fresh install, and had the same results.
 
The only other stand-out data point I've found is a WARNING in syslog when I restart the webcit service:
 
Jun 11 11:45:55 KRILLAN webcit[2492]: WARNING: getpwuid(65534): Success#012Group IDs will be incorrect.
Jun 11 11:45:55 KRILLAN webcit[2492]: Changing uid to 65534
Jun 11 11:45:55 KRILLAN webcit[2495]: Listening on socket 7
Jun 11 11:45:55 KRILLAN webcit[2495]: Requesting cipher list: DEFAULT
Jun 11 11:45:55 KRILLAN webcit[2495]: WARNING: getpwuid(65534): File exists#012Group IDs will be incorrect.
Jun 11 11:45:55 KRILLAN webcit[2495]: Changing uid to 65534
 
I've found a reference to this error on https://www.mail-archive.com/room_citadel_development@uncensored.citadel.org/msg09459.html, which I'm going to do when I have a moment, though I'm not sure if this is a relevant issue.
 
 
 
If anyone is interested in the syslog, let me know and I'll email it over (since this forums doesn't seem to support code boxes or post attachments). Again, I'd love suggestions on how to proceed since I'm at a bit of a brick wall, but past anything else I'd also be interested in suggestions on an optimal distro of linux to run this on, as on Saturday, if I'm still stuck I'm going to go nuclear option on this and start from scratch with something else than Debian Jessie.


[#] Thu Jun 11 2015 13:44:48 EDT from CarlGambolputty @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Ugh. Ignore the double, please. Apparently hitting the pound sign while investing how to edit a post makes a copy somehow.



[#] Fri Jun 12 2015 11:14:52 EDT from CarlGambolputty @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Quick update - solving the getpwid bug didn't change anything, unfortunately. I'm running out of time to get this rolling, so I'm going to run this on Ubuntu Server 15.04 and giving up on Debian's new release for now. Will post back with results.



[#] Sun Jun 14 2015 12:18:20 EDT from markus @ Uncensored

Subject: Re: Again: Citadel cannot be connected after some time

[Reply] [ReplyQuoted] [Headers] [Print]

Hello again,

 

Well, I did some research. According to my log files,  my citadel server was quite often visited by a computer x with the domain <x>.compute-1.amazonawa.com. This computer tries to connect my citadel server via imap.

And it seams to me that every time this takes place citadel gets inaccessable for any of my mail clients including webcite.

 

Does anybody has any suggetion to avoid this behavoiur of citadel?

 

Thanks for your replay.

 

Markus

 

 

 

 

 

 

 

 

Wed Jun 10 2015 07:37:39 EDT from markus @ Uncensored Subject: Again: Citadel cannot be connected after some time

Hello again,

 

does anybody have any idea to my problem posted bellow? If more or more precise information is required, l'm glad about any question.

 

Thank you.

 

Markus

 

 

 

 

Sun May 10 2015 18:16:28 EDT from markus @ Uncensored Subject: Citadel cannot be connected after some time

Hello Citadel Support,

 

I'm running Citadel on Debian 7.

My problem is, that Citadel runs for some hours without any obvious problem (mails getting delivered both ways) and then Citadel gets inaccessable by any email client that I use (WebCit, Thunderbird and TypeMail). When I restart citadel, it runs again without problems and all clients can connect, but after a couple of hours the cycle starts over again.

I have already exported and reimported my database in order to be sure it is not a problem with a corrupted database. But that seems not be the case, because the problem was still there.

Does anybody has any idea what's going on? And: Where can I find usefull log files?

 

Thank you very much in advance!

Markus

 

 



 



 



[#] Mon Jun 15 2015 10:42:28 EDT from IGnatius T Foobar @ Uncensored

Subject: Re: Again: Citadel cannot be connected after some time

[Reply] [ReplyQuoted] [Headers] [Print]

Well, I did some research. According to my log files,  my citadel
server was quite often visited by a computer x with the domain
<x>.compute-1.amazonawa.com. This computer tries to connect my
citadel server via imap.

There's all sorts of bad behavior on AWS. Unfortunately you are simply being DoSed, and you just need to block that IP address using iptables or your network's firewall.

[#] Mon Jun 15 2015 11:36:44 EDT from CarlGambolputty @ Uncensored

Subject: Re: Suggestions for migrating Exchange/SBS 2011 to Citadel/Debian Jessie

[Reply] [ReplyQuoted] [Headers] [Print]

I've finally solved my problem. It's a very stupid and silly-simple solution.

After getting Ubuntu Server up and running with citadel install with all dependencies, I encountered the same exact problem as before. Same messages and everything. So I started throwing crap at the wall for my citadel server configuration values, just to see if something would stick. 

Turns out that Citadel does not like using the IP address of your DC; instead, it wants the complete DC.domain.local, where DC is the domain controller server name, and domain.local is your domain name. Unfortunately, the documentation for integration with Windows AD tells you to use the IP address OR hostname. And silly me, I thought 'why bother using DNS when I can just specify the logical address?' Silly, silly me. I recommend that the documentation be upgraded to reflect this for Windows SBS 2011 (not sure if this happens on other Windows server distros).

Now I have to move everything over, but at least I have single sign-on working. 



[#] Mon Jun 15 2015 23:54:56 EDT from IGnatius T Foobar @ Uncensored

Subject: Re: Suggestions for migrating Exchange/SBS 2011 to Citadel/Debian Jessie

[Reply] [ReplyQuoted] [Headers] [Print]

Interesting. I've never heard of that happening before, but I'm glad you found the problem.

[#] Tue Jun 16 2015 04:55:37 EDT from itisme @ Uncensored

Subject: 9.01 Release

[Reply] [ReplyQuoted] [Headers] [Print]

Heyho, I am running citadel 8.24 and I am thinking of updating to 9.01 if following things are true:

 

Is it true that 9.01 has an read-only usenet server?

Does 9.01 have CalDAV CardDAV support? (the Development roadmap says so: www.citadel.org/doku.php/documentation:roadmap#development.roadmap)

 

Thabks in Advance!



[#] Tue Jun 16 2015 10:14:15 EDT from IGnatius T Foobar @ Uncensored

Subject: Re: 9.01 Release

[Reply] [ReplyQuoted] [Headers] [Print]


Yikes, that roadmap is way out of date.

However, the current production release of Citadel *does* have a read-only NNTP server.

[#] Tue Jun 16 2015 14:26:07 EDT from itisme @ Uncensored

Subject: Re: 9.01 Release

[Reply] [ReplyQuoted] [Headers] [Print]

Does it also have CalDAV/CardDAV support?



[#] Tue Jun 16 2015 14:37:30 EDT from itisme @ Uncensored

Subject: Re: 9.01 Release

[Reply] [ReplyQuoted] [Headers] [Print]

And What do you mean with "production release"?



[#] Wed Jun 17 2015 06:55:14 EDT from RoddyJoff @ Uncensored

Subject: Accessing Citadel from Apple Mail + Multiple Domain Configuration

[Reply] [ReplyQuoted] [Headers] [Print]

Hello,

I'm a new user of Citadel so please treat me gently ;-)

 

I have installed Citadel on a RaspberryPI and can send and receive mail from it and also to it from an external account.

 

I have two questions which I hope someone can help me with.

1 - How do I configure users to use different domains? If I enter multiple domain name info I do not have the option to allocate a domain when creating a user and the user is allocated the first domain I configure.

2 - Does anyone know how to configure Apple Mail to use the Citadel server?

 

Many thanks,

RJ

 



[#] Thu Jun 18 2015 15:15:03 EDT from IGnatius T Foobar @ Uncensored

Subject: Re: 9.01 Release

[Reply] [ReplyQuoted] [Headers] [Print]

The current production release is 9.01. It does have the read-only NNTP server but it does not yet have CalDAV/CardDAV.

[#] Thu Jun 18 2015 20:38:00 EDT from platonov @ Uncensored

Subject: Re: 9.01 Release citserver crash

[Reply] [ReplyQuoted] [Headers] [Print]

I just installed 9.01 on Ubuntu Ubuntu 14.04.2 LTS.

Looks like citserver is crashing and restarting all the time.

Here's what I found in /var/log/messages:

Jun 19 03:16:26 preciseinfo kernel: citserver[8800]: segfault at 59c ip 00000000f74fd60a sp 00000000f69712b0 error 4 in libcitadel.so.4.0.901  f74f1000+22000]

By spending a couple of hours struggling, it seems that some things do work at least for a couple of commands. I was able to login on main admin account and was able to do some things. But it always ends up saying that it can not connect and stay connected to citserver.

Any idea?





[#] Thu Jun 18 2015 23:12:39 EDT from IGnatius T Foobar @ Uncensored

Subject: Re: 9.01 Release citserver crash

[Reply] [ReplyQuoted] [Headers] [Print]

If you can get citserver to produce a core dump, that dump plus the executable and gdb can give us a stack trace. That will allow us to zero in on the problem.

[#] Fri Jun 19 2015 05:11:20 EDT from platonov @ Uncensored

Subject: Re: 9.01 Release citserver crash

[Reply] [ReplyQuoted] [Headers] [Print]

Anyone can provide more detailed instructions to get this done?

Basically, it is a headless server situation and I was doing the upgrade from 8.24.

What seems to be strange is that easyinstall does not seem to terminate properly. Here's its last messages:

  • Please enter the name of the Citadel user account that should be
    granted administrative privileges once created.

    Citadel Easy Install is aborting.

    A log file has been written to /tmp/citadel-install-log.txt
    Reading this file may tell you what went wrong. If you
    need to ask for help on the support forum, please post the
    last screenful of text from this log.

And here is what I see in that log:

  • --2015-06-19 12:03:46--  http://easyinstall.citadel.org/citadel-init-scripts.tar
  • Resolving easyinstall.citadel.org (easyinstall.citadel.org)... 216.150.130.115
  • Connecting to easyinstall.citadel.org (easyinstall.citadel.org)|216.150.130.115|:80... connected.
  • HTTP request sent, awaiting response... 200 OK
  • Length: 10240 (10K) [application/x-tar]
  • Saving to: 'citadel-init-scripts.tar'
  •      0K ..........                                            100% 71.7K=0.1s
  • 2015-06-19 12:03:47 (71.7 KB/s) - 'citadel-init-scripts.tar' saved [10240/10240]
  • Operating system: Linux Debian jessie/sid ( 4.0.5-x86_64-linode58 x86_64)
Thu Jun 18 2015 23:12:39 EDT from IGnatius T Foobar @ Uncensored Subject: Re: 9.01 Release citserver crash
If you can get citserver to produce a core dump, that dump plus the executable and gdb can give us a stack trace. That will allow us to zero in on the problem.

 



[#] Fri Jun 19 2015 05:29:31 EDT from platonov @ Uncensored

Subject: Re: 9.01 Release citserver crash

[Reply] [ReplyQuoted] [Headers] [Print]

Forgot to mention that as soon as I enter "user account that should be granted administrative privileges", easyinstall terminates immediately without showing a single step of progress. At that point the citadel server starts running and then immediately crashing and so on.

Fri Jun 19 2015 05:11:20 EDT from platonov @ Uncensored Subject: Re: 9.01 Release citserver crash

Anyone can provide more detailed instructions to get this done?

Basically, it is a headless server situation and I was doing the upgrade from 8.24.

What seems to be strange is that easyinstall does not seem to terminate properly. Here's its last messages:

  • Please enter the name of the Citadel user account that should be
    granted administrative privileges once created.

    Citadel Easy Install is aborting.

    A log file has been written to /tmp/citadel-install-log.txt
    Reading this file may tell you what went wrong. If you
    need to ask for help on the support forum, please post the
    last screenful of text from this log.

And here is what I see in that log:

  • --2015-06-19 12:03:46--  http://easyinstall.citadel.org/citadel-init-scripts.tar
  • Resolving easyinstall.citadel.org (easyinstall.citadel.org)... 216.150.130.115
  • Connecting to easyinstall.citadel.org (easyinstall.citadel.org)|216.150.130.115|:80... connected.
  • HTTP request sent, awaiting response... 200 OK
  • Length: 10240 (10K) [application/x-tar]
  • Saving to: 'citadel-init-scripts.tar'
  •      0K ..........                                            100% 71.7K=0.1s
  • 2015-06-19 12:03:47 (71.7 KB/s) - 'citadel-init-scripts.tar' saved [10240/10240]
  • Operating system: Linux Debian jessie/sid ( 4.0.5-x86_64-linode58 x86_64)
Thu Jun 18 2015 23:12:39 EDT from IGnatius T Foobar @ Uncensored Subject: Re: 9.01 Release citserver crash
If you can get citserver to produce a core dump, that dump plus the executable and gdb can give us a stack trace. That will allow us to zero in on the problem.

 



 



[#] Fri Jun 19 2015 06:48:33 EDT from dothebart @ Uncensored

Subject: Re: 9.01 Release citserver crash

[Reply] [ReplyQuoted] [Headers] [Print]

Could you have a look at the syslog, and see what its printing there before it goes away?



Go to page: First ... 49 50 51 52 [53] 54 55 56 57 ... Last