Show Posts
This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.
Pages: [1] 2
1
Installation and Upgrades / [Q] Lost ability to edit zarafa-rules after apt-get autoremove
« on: August 13, 2012, 12:25:49 am »
I did something stupid the other day, ran apt-get autoremove. Didn't seem to harm anything at first, but users have identified an issue where they can't edit the zarafa rule-list anymore, it stops at "Loading...".
I've tried uninstalling and re-installing Groupware (leads to the webinterface not being able to connect to zarafa, after reboot as well), tried re-installing fckeditor (which is the only item I specifically can remember being uninstalled by apt-get autoremove), didn't help.
Does anybody know which packages I need to install to get rule-editing working again?
::Trym
I've tried uninstalling and re-installing Groupware (leads to the webinterface not being able to connect to zarafa, after reboot as well), tried re-installing fckeditor (which is the only item I specifically can remember being uninstalled by apt-get autoremove), didn't help.
Does anybody know which packages I need to install to get rule-editing working again?
::Trym
2
Installation and Upgrades / [SOLVED] Mail not working after migrating 2.0->2.2
« on: June 07, 2012, 11:19:35 pm »
Hi guys. For the third time this year I tried migrating from 2.0 to 2.2. This leads to mail failing.
The components are running, but queue-management/status shows
"connect to 127.0.0.1[127.0.0.1]:2003:Connection refused." for all mails.
I figured this would be fixed in time, so I've waited a few months between each attempt.
Does anybody know what I can do before/after migration to get this fixed? (I've searched for similar questions, and all I find are single-post questions with no answers or follow-ups.)
The components are running, but queue-management/status shows
"connect to 127.0.0.1[127.0.0.1]:2003:Connection refused." for all mails.
I figured this would be fixed in time, so I've waited a few months between each attempt.
Does anybody know what I can do before/after migration to get this fixed? (I've searched for similar questions, and all I find are single-post questions with no answers or follow-ups.)
3
Installation and Upgrades / [SOLVED] UserCorner external e-mail malfunction after migration.
« on: November 29, 2011, 01:19:24 am »
I've in vain tried to get external e-mail in usercorner to work after migrating to 2.2.
As part of a process I'm doing anyway I've virtualised the physical machines, re-configured the network and then tried to migrate. The migration itself is succesful, but after installing user-corner and enabling it (since it's now a separate module) it gives an error message when trying to access the "External e-mail retreival" page. No external mail gets fetched either. I've gone back to the original 2.0-image and tried various combinations of ticking and unticking options in different modules before the migration, always with the same outcome.
The logs give no info I can relate to this error, the usercorner module is not reporting any errors I can find.
Does anybody have any clue? Purging existing data for usercorner is not an option.
As part of a process I'm doing anyway I've virtualised the physical machines, re-configured the network and then tried to migrate. The migration itself is succesful, but after installing user-corner and enabling it (since it's now a separate module) it gives an error message when trying to access the "External e-mail retreival" page. No external mail gets fetched either. I've gone back to the original 2.0-image and tried various combinations of ticking and unticking options in different modules before the migration, always with the same outcome.
The logs give no info I can relate to this error, the usercorner module is not reporting any errors I can find.
Does anybody have any clue? Purging existing data for usercorner is not an option.
4
Installation and Upgrades / [SOLVED] Z-push -> Android 2.2.2
« on: March 20, 2011, 02:01:35 pm »
Got a shiny new toy yesterday, an Optimus 2x running Android 2.2.2 
First thing I wanted to try was of course syncing with Zarafa using the built in Android client (which is called "MS Exchange", although it's an ActiveSync app{*}), as this would be immensely useful for me. The net seems to be full of success stories (as well as people wanting to smash their phones), so I'm figuring I'm doing something wrong.
I cannot get past the "Connecting to server - The phone is connecting to the email-server"-bit. It just sits there forever rotating its little "wait a while" circle.
While the phone is "connecting" there definitively is something going on; apache, mysql and zarafa-server each use some processor time, seemingly in a repeating pattern.
I'm connecting to the "domain", ie. virtual web-server address which Zarafa is set at. If I connect to the same "<domain>/Microsoft-Server-ActiveSync" in a browser I get challenged by a username/password auth-box, that bit seems to be working for both http and https. I've tried connecting with SSL set to both on and off.
I figured maybe my mailbox was too large and something timed out, so I made another quick setup in a virtual machine with an empty account, same thing happens.
Also, if I type a non-existing server address or an address without Z-push listening, the app doesn't even get past the point of configuring the connection, so it does find something it likes before trying to connect.
I don't know where to go from here, the usual logs show no clues.
The parameters you can input in the client are:
e-mail address:
username:
password
domain:
Server address:
The only mysterious parameter is the domain, I'm not sure what to input, I've tried a blank domain as well as the zarafa-domain and the virtual e-mail domain, it doesn't seem to make any difference, most guides leave it blank.
If I had to guess, I'd guess that something goes wrong during the authentication-process, and that it just keeps trying (looking at the processes in 'top'.) I've tried with and without a certificate for Zarafa Gateway Server, which shouldn't really matter. (Wrong certificate perhaps? I've no clue how to make a correct one if that's the case.)
Other thoughts:
*1) I may have missed the plot completely, and the "MS Exchange" listed under Sync is not an ActiveSync client, but a full Outlook Client, in which case I need to install the community version of Zarafa. There's supposed to be built-in support for Active Sync in Froyo though, so if it's the wrong app, where the heck is the correct one?
2) The security settings on the phone does not match the criteria for ActiveSync, and the phone somehow neglects to notice/tell me.
3) Z-push in Zentyal's Zarafa is not correctly configured?
4) Z-push version is too old?
Any ideas at all welcome, feel free to speculate. If you've got any of this working, or even just semi-working I'd be very thankful for any configuration pointers.
::Trym
First thing I wanted to try was of course syncing with Zarafa using the built in Android client (which is called "MS Exchange", although it's an ActiveSync app{*}), as this would be immensely useful for me. The net seems to be full of success stories (as well as people wanting to smash their phones), so I'm figuring I'm doing something wrong.
I cannot get past the "Connecting to server - The phone is connecting to the email-server"-bit. It just sits there forever rotating its little "wait a while" circle.
While the phone is "connecting" there definitively is something going on; apache, mysql and zarafa-server each use some processor time, seemingly in a repeating pattern.
I'm connecting to the "domain", ie. virtual web-server address which Zarafa is set at. If I connect to the same "<domain>/Microsoft-Server-ActiveSync" in a browser I get challenged by a username/password auth-box, that bit seems to be working for both http and https. I've tried connecting with SSL set to both on and off.
I figured maybe my mailbox was too large and something timed out, so I made another quick setup in a virtual machine with an empty account, same thing happens.
Also, if I type a non-existing server address or an address without Z-push listening, the app doesn't even get past the point of configuring the connection, so it does find something it likes before trying to connect.
I don't know where to go from here, the usual logs show no clues.
The parameters you can input in the client are:
e-mail address:
username:
password
domain:
Server address:
The only mysterious parameter is the domain, I'm not sure what to input, I've tried a blank domain as well as the zarafa-domain and the virtual e-mail domain, it doesn't seem to make any difference, most guides leave it blank.
If I had to guess, I'd guess that something goes wrong during the authentication-process, and that it just keeps trying (looking at the processes in 'top'.) I've tried with and without a certificate for Zarafa Gateway Server, which shouldn't really matter. (Wrong certificate perhaps? I've no clue how to make a correct one if that's the case.)
Other thoughts:
*1) I may have missed the plot completely, and the "MS Exchange" listed under Sync is not an ActiveSync client, but a full Outlook Client, in which case I need to install the community version of Zarafa. There's supposed to be built-in support for Active Sync in Froyo though, so if it's the wrong app, where the heck is the correct one?
2) The security settings on the phone does not match the criteria for ActiveSync, and the phone somehow neglects to notice/tell me.
3) Z-push in Zentyal's Zarafa is not correctly configured?
4) Z-push version is too old?
Any ideas at all welcome, feel free to speculate. If you've got any of this working, or even just semi-working I'd be very thankful for any configuration pointers.
::Trym
5
Installation and Upgrades / Log full of useless debug messages.
« on: March 17, 2011, 09:50:50 pm »
For the past week or so, maybe longer, I've been getting thousands upon thousands of these entries in /var/log/ebox/ebox.log:
------------
2011/03/17 21:17:04 DEBUG> PgDBEngine.pm:328 EBox::PgDBEngine::query - Error querying data: SELECT rfc931 AS username, remotehost AS ip, domain_from_url(url) AS domain, event, code, SUM(bytes) AS bytes, COUNT(event) AS hits FROM squid_access WHERE (squid_access.timestamp >= '2011-3-1 00:00:00' AND squid_access.timestamp < date '2011-3-01 00:00:00' + interval '1 month') GROUP BY username, ip, domain, event, code ; , ERROR: function domain_from_url(character varying) does not exist
LINE 1: SELECT rfc931 AS username, remotehost AS ip, domain_from_url...
^
HINT: No function matches the given name and argument types. You might need to add explicit type casts.
------------
I have HTTP Proxy logging turned OFF.
Now these appear to be DEBUG messages, and nothing I'm aware of is malfunctioning. It's just very annoying to hunt through thousands of useless lines of log when looking for clues.
Is there a DEBUG-flag somewhere accidentally turned on - in which case - can I switch it off?
Any insight welcome.
::Trym
------------
2011/03/17 21:17:04 DEBUG> PgDBEngine.pm:328 EBox::PgDBEngine::query - Error querying data: SELECT rfc931 AS username, remotehost AS ip, domain_from_url(url) AS domain, event, code, SUM(bytes) AS bytes, COUNT(event) AS hits FROM squid_access WHERE (squid_access.timestamp >= '2011-3-1 00:00:00' AND squid_access.timestamp < date '2011-3-01 00:00:00' + interval '1 month') GROUP BY username, ip, domain, event, code ; , ERROR: function domain_from_url(character varying) does not exist
LINE 1: SELECT rfc931 AS username, remotehost AS ip, domain_from_url...
^
HINT: No function matches the given name and argument types. You might need to add explicit type casts.
------------
I have HTTP Proxy logging turned OFF.
Now these appear to be DEBUG messages, and nothing I'm aware of is malfunctioning. It's just very annoying to hunt through thousands of useless lines of log when looking for clues.
Is there a DEBUG-flag somewhere accidentally turned on - in which case - can I switch it off?
Any insight welcome.
::Trym
6
Installation and Upgrades / 2.1 User quota location
« on: March 17, 2011, 05:31:04 pm »
Installed a fresh 2.1 which (finally) let me create users without errors.
I immediately noticed that the first user had been created with a quota of 100MB (why? Seems the default should be "no quota".)
Went to change the quato for the user. Turns out you cannot do that without resetting the user's password, as the quato settings is bundled in with the full name and password settings.
Also, after settting the quota to 0 and and replacing the password, the "User quota" field suddenly displays the text "extra_passwords."
Not exactly a bug, but neither useful or pretty.
EDIT: I actually don't mind the location of the quota setting, what I do mind are the password fields. You should be able to change both the full names, comments and quota without changing the password. I think you should replace/remove the password fields and instead add a button for the specific intent of changing the password, and in other cases just leave it untouched. This applies to 2.0.x as well.
::Trym
I immediately noticed that the first user had been created with a quota of 100MB (why? Seems the default should be "no quota".)
Went to change the quato for the user. Turns out you cannot do that without resetting the user's password, as the quato settings is bundled in with the full name and password settings.
Also, after settting the quota to 0 and and replacing the password, the "User quota" field suddenly displays the text "extra_passwords."
Not exactly a bug, but neither useful or pretty.
EDIT: I actually don't mind the location of the quota setting, what I do mind are the password fields. You should be able to change both the full names, comments and quota without changing the password. I think you should replace/remove the password fields and instead add a button for the specific intent of changing the password, and in other cases just leave it untouched. This applies to 2.0.x as well.
::Trym
7
Installation and Upgrades / [SOLVED] Mail: temporary error 0x80040116
« on: March 05, 2011, 10:59:48 am »
Today I woke up to a lot of these messages in the email-log: "temporary failure. Command output: Unable to commit message: 0x80040116 Unable to deliver message to 'trym', error code: 0x80040116". (Status:Deferred)
The affected mails are still in the mail-queue. There are of course a lot of mails to other users as well, same error codes.
I've restarted the mail-module, the mail-filter module, and when that didn't work I rebooted the whole thing. Didn't help.
Does anybody have any clue what can be done about this?
::Trym
The affected mails are still in the mail-queue. There are of course a lot of mails to other users as well, same error codes.
I've restarted the mail-module, the mail-filter module, and when that didn't work I rebooted the whole thing. Didn't help.
Does anybody have any clue what can be done about this?
::Trym
8
Installation and Upgrades / [SOLVED?] How to auto-reboot after fsck?
« on: February 24, 2011, 09:33:38 pm »
It's the windy season, and predictably there's at least one power-cut a week. Although I'm personally going to invest in some form of UPS, there's no way I can get the money for it at work, the servers are basically built out of scrap parts. I can convince them to invest in SSDs for the servers as that has an immediate, tangible benefit.
I have a problem with the first boot after a power-cut using an SSD (both on the official kernel and on kernel 2.6.34.)
Simply put, fsck starts checking/repairing the file-system, but the system loads so much more quickly from the SSD that slapd tries to start too early, fails, and the entire system hangs a little later in the startup-procedure (after starting the DNS-module.)
Doing a ctrl-alt-delete solves it, the system reboots, and everything works fine on the next startup.
Machines using traditional hard-disks don't have this problem, they reboot fine after the power-cuts.
So my question is: What is the easiest way to force a reboot if fsck has made corrections? I still don't have any intimate knowledge of the ubuntu/zentyal startup-scripts, so would be very thankful for any pointers.
::Trym
I have a problem with the first boot after a power-cut using an SSD (both on the official kernel and on kernel 2.6.34.)
Simply put, fsck starts checking/repairing the file-system, but the system loads so much more quickly from the SSD that slapd tries to start too early, fails, and the entire system hangs a little later in the startup-procedure (after starting the DNS-module.)
Doing a ctrl-alt-delete solves it, the system reboots, and everything works fine on the next startup.
Machines using traditional hard-disks don't have this problem, they reboot fine after the power-cuts.
So my question is: What is the easiest way to force a reboot if fsck has made corrections? I still don't have any intimate knowledge of the ubuntu/zentyal startup-scripts, so would be very thankful for any pointers.
::Trym
9
Installation and Upgrades / Seeking advice for TRIM on Zentyal
« on: February 16, 2011, 01:38:10 pm »
I recently purchased a new SSD for my desktop, with the specific intent of replacing the Zentyal hard-drive with my "old" SSD.
I did, installed a fresh Zentyal 2.0.3 (no LVM, just plain ext4), restored the system from a backup and it works like a charm, Zarafa especially is a lot more responsive, as expected.
However...
I'd read all over the place that Linux has full TRIM support. I mistakenly assumed that 10.04 LTS certainly would have it. It turns out that it is only fully supported from kernel 2.6.33, whereas 10.04 LTS will always be stuck at 2.6.32.
In a virtual machine I've upgraded 2.6.32 to 2.6.33 without problems, the only visible change so far is that the Zentyal splash-screen gets replaced with the ubuntu splash-screen, not a problem. Everything else seems to work.
I dare not do this on a production system before seeking advice from any gurus out there. I have no clue which other changes there have been in 32>33, and if any of them are likely to break any component of Zentyal. Do you?
There won't be another LTS until april 2012, so I would like to strongly encourage Zentyal to either:
...whichever fits the best. Anyone with an eye on performance will want to run Zentyal from an SSD, and it can only be beneficial for Zentyal to say you fully support it.
::Trym
I did, installed a fresh Zentyal 2.0.3 (no LVM, just plain ext4), restored the system from a backup and it works like a charm, Zarafa especially is a lot more responsive, as expected.
However...
I'd read all over the place that Linux has full TRIM support. I mistakenly assumed that 10.04 LTS certainly would have it. It turns out that it is only fully supported from kernel 2.6.33, whereas 10.04 LTS will always be stuck at 2.6.32.
In a virtual machine I've upgraded 2.6.32 to 2.6.33 without problems, the only visible change so far is that the Zentyal splash-screen gets replaced with the ubuntu splash-screen, not a problem. Everything else seems to work.
I dare not do this on a production system before seeking advice from any gurus out there. I have no clue which other changes there have been in 32>33, and if any of them are likely to break any component of Zentyal. Do you?
There won't be another LTS until april 2012, so I would like to strongly encourage Zentyal to either:
- Backport full TRIM-support.
- Upgrade the kernel.
- Create a package in the Zentyal repository which, if an SSD is detected, will install a newer version of hdparm than the Ubuntu repository, along with its wiper.sh script, and set up a cronjob to run it.
...whichever fits the best. Anyone with an eye on performance will want to run Zentyal from an SSD, and it can only be beneficial for Zentyal to say you fully support it.
::Trym
10
Installation and Upgrades / [SOLVED] How to correctly report bugs?
« on: February 13, 2011, 05:41:42 pm »
As some of you may know, I've been working on a how-to for backup and restoration of a Zentyal server. During the course I've found dozens and dozens of small bugs.
The usual routine for other types of software is to first report and discuss a possible bug in the forum, and then if there's a consensus it's a bug, create a ticket in trac.
This doesn't seem to be how it works in the Zentyal forums.
Can anybody tell me the correct procedure for discussing/reporting bugs?
::Trym
The usual routine for other types of software is to first report and discuss a possible bug in the forum, and then if there's a consensus it's a bug, create a ticket in trac.
This doesn't seem to be how it works in the Zentyal forums.
Can anybody tell me the correct procedure for discussing/reporting bugs?
::Trym
11
Installation and Upgrades / [SOLVED] Zarafa HTML-mail no longer working
« on: February 07, 2011, 02:42:33 pm »
I've recently lost the ability to write HTML-formatted mails in Zarafa. Usually I have it set to use text-format, but a few days ago I need to write a mail with some formatting. Went to preferences, set "Compose mail in this format:" to HTML.
I've fiddled to and from with it the last few days, tried different browsers, reset all webbrowser profiles and plugins, set the "Compose mail..." to text and back, nothing I do seems to help.
The "New Mail" window always shows mail as pure text, without any buttons for formatting.
Has anybody else experienced this?
::Trym
I've fiddled to and from with it the last few days, tried different browsers, reset all webbrowser profiles and plugins, set the "Compose mail..." to text and back, nothing I do seems to help.
The "New Mail" window always shows mail as pure text, without any buttons for formatting.
Has anybody else experienced this?
::Trym
12
Installation and Upgrades / Port-forwarding and transparent HTTP-Proxy.
« on: February 04, 2011, 02:54:23 pm »
Zentyal 2.0 fully updated, one external nic PPPoE connected to modem, one internal nic connected to a switch, clients connected to switch.
I cannot get port-forwarding to work while using transparent proxy, no matter which firewall-rules I add.
If I disable transparent proxying, port-forwarding works as expected, without the use of any additional firewall-rules.
Trac mentions this as a bug which has been marked as solved.
Is this a regression, or am I just missing some simple configuration-trick?
::Trym
I cannot get port-forwarding to work while using transparent proxy, no matter which firewall-rules I add.
If I disable transparent proxying, port-forwarding works as expected, without the use of any additional firewall-rules.
Trac mentions this as a bug which has been marked as solved.
Is this a regression, or am I just missing some simple configuration-trick?
::Trym
13
Installation and Upgrades / [SOLVED] Cannot create VPN. Can't use an undefined value as a HASH reference.
« on: January 18, 2011, 04:55:21 pm »
"Can't use an undefined value as a HASH reference at /usr/share/perl5/EBox/OpenVPN/Model/Servers.pm line 394."
That is the error message I get on two different physical servers, originally Zentyal 2.02 installations, fully upgraded, one 32-bit, one 64-bit.
The error page instructs you to click on 'here' and 'here', which is supposed to get a log-file and link to trac, neither of which work (page not found.)
I'm using no non-english characters in any of the certificates or vpn-site names, no dashes, no spaces, just letters and a dot. Both servers are serving real domain-names, and every external component has a certificate installed.
I've searched forums and trac, and the relevant posts states that this message was from a bug dealing with foreign characters, which in turn got fixed.
Doing a manual dump of /var/log/ebox/ebox.log, the last 7 lines show:
It claims there is no CA certificate, which there certainly is.
If anybody has even the slighest idea of anything I can do to get VPN up and running, please respond. VPN was a heavy selling point, I need to get it going and not look like a fool for recommending Zentyal.
::Trym
That is the error message I get on two different physical servers, originally Zentyal 2.02 installations, fully upgraded, one 32-bit, one 64-bit.
The error page instructs you to click on 'here' and 'here', which is supposed to get a log-file and link to trac, neither of which work (page not found.)
I'm using no non-english characters in any of the certificates or vpn-site names, no dashes, no spaces, just letters and a dot. Both servers are serving real domain-names, and every external component has a certificate installed.
I've searched forums and trac, and the relevant posts states that this message was from a bug dealing with foreign characters, which in turn got fixed.
Doing a manual dump of /var/log/ebox/ebox.log, the last 7 lines show:
Code: [Select]
2011/01/18 18:09:37 INFO> Global.pm:535 EBox::Global::saveAllModules - Saving config and restarting services: firewall openvpn logs
2011/01/18 18:09:37 INFO> Base.pm:151 EBox::Module::Base::save - Restarting service for module: firewall
2011/01/18 18:09:48 INFO> Base.pm:773 EBox::Module::Base::_hook - Running hook: /etc/ebox/hooks/firewall.postservice 1
2011/01/18 18:09:49 INFO> Base.pm:151 EBox::Module::Base::save - Restarting service for module: openvpn
2011/01/18 18:09:49 ERROR> Server.pm:201 EBox::OpenVPN::Server::caCertificatePath - No CA certificate
2011/01/18 18:09:49 INFO> Base.pm:151 EBox::Module::Base::save - Restarting service for module: logs
2011/01/18 18:09:50 ERROR> Global.pm:632 EBox::Global::saveAllModules - The following modules failed while saving their changes, their state is unknown: openvpn
It claims there is no CA certificate, which there certainly is.
If anybody has even the slighest idea of anything I can do to get VPN up and running, please respond. VPN was a heavy selling point, I need to get it going and not look like a fool for recommending Zentyal.
::Trym
14
Installation and Upgrades / PPPoE does not reconnect after ADSL downtime.
« on: January 15, 2011, 01:19:21 pm »
This is bugging me to no end. Twice during the last three weeks I've had to walk through a blizzard simply to get a Zentyal server back online after the ADSL-connection went down for a short time.
Although this is a HUGE BUG, that's not the purpose of this post, the purpose is to find a solution I can reliably use.
I'm not much of a script-writer (at least not yet), so I'm hoping someone out there has solved this already.
The script will, I guess, run as a cronjob (doesn't have to run more than every 10 minutes or so for my needs) which checks that it can ping the standard gateway, and if it can't it'll try to reconnect. (Either by sending ppp-connect commands, or by bringing the interface down, then back up.)
Does anybody know of such a script for Zentyal?
::Trym
Although this is a HUGE BUG, that's not the purpose of this post, the purpose is to find a solution I can reliably use.
I'm not much of a script-writer (at least not yet), so I'm hoping someone out there has solved this already.
The script will, I guess, run as a cronjob (doesn't have to run more than every 10 minutes or so for my needs) which checks that it can ping the standard gateway, and if it can't it'll try to reconnect. (Either by sending ppp-connect commands, or by bringing the interface down, then back up.)
Does anybody know of such a script for Zentyal?
::Trym
15
Installation and Upgrades / [SOLVED] How do I save changes from the command line?
« on: January 09, 2011, 03:24:32 pm »
Title says it all, I'm looking for the command/script which will save the current changes to the configuration from the command-line. I've written a HOW-TO on disaster recovery, and I'm happy with it as it is, but having to switch between a shell and firefox constantly to save settings is not very elegant (and not always safe, as there's a mismatch between the cookies and the actual changes.)
Does anybody know?
::Trym
Does anybody know?
::Trym
Pages: [1] 2