Language:
switch to room list switch to menu My folders
Go to page: First ... 29 30 31 32 [33] 34 35
[#] Wed Nov 15 2017 12:55:11 EST from warbaby @ Uncensored

Subject: Re: howto pass blocked port 25

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

Like others have stated, you need an SMTP  provider.  

We're using sendgrid (sendgrid.com) as a smarthost from one box, 30 days free.. then drops down to 100 email "development" level, which is plenty good for what we're doing. There are many others designed basically to deliver email newsletters.  Sendgrid does a decent job with delivery.   Haven't had anything turn up in a spam folder yet, & we send to Government Mail (Gmail) all the time. 

 

 

Sun Nov 12 2017 06:05:17 PM EST from wildbill @ Uncensored Subject: howto pass blocked port 25

My ISP blocks port 25. I cannot persuade citadel to use any other.  I have tried disabling the SMTP port (-1) and tried setting it to 587. For both of these Wireshark tells me citadel is still trying to send on port 25. How can I change it to send on port 587 or 465?

 



 



[#] Sun Nov 19 2017 14:07:04 EST from joseporti @ Uncensored

Subject: Broken download vmware easy install link

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

Hello everyone,

I would like to ask you if you can upload the following link again:

http://easyinstall.citadel.org/citadel-vmware-appliance-2012.zip

It's broken and I need it, please. I will be very helpful. Thank you Regards



[#] Mon Nov 20 2017 04:05:41 EST from techacq1Aa @ Uncensored

Subject: "Outbound Queue" invisible bounce sender to flood the database

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

Server crashed Sunday with a full disk.

First suspicion was a hack but that was not the case.

The database had swollen to fit the space allotted and promptly died.

The error logs then suffocated the remaining system.

Examining the /data/*.log files showed massive graphics being shovelled through,

 repeatedly. As if someone had fired up an endless image regurgitation device.

The var/logs were showing the system trying to endlessly and without hesitation

 to send this 4MB message over and over again to a bounce address.

 

Our loveliest user ever (very nice lady 'everyone loves aunty')

had sent an 4mB foto email with a bounce-back notification enabled

to a domain that no longer exists.

We had no idea this could be a problem

 

The bounceback has no subject and no sender

so when looking in "View the Outbound SMTP queue" which defaulted to view 11

as if by magic nothing appears.

While the database slaps another 4mB on every minute or so.

Devil was in the details trying to identify what the issue was

but the logs showed this undeliverable email trying to resend all weekend long

 until /usr/local/citadel/data finally gasped its last breath.

Hosing out the database directory with last weeks stale image,

helped for about 30seconds when the system resumed its suicidal bent.

 

A mail that cannot be seen in the outbound queue ?!

Only after wailing and gnashing of teeth

did the idea of changing the view show the culprit.

changing the view from "11" to "1" displayed the cryptic line which could then be deleted.

Then the regurgitation ended and peace flooded the server room once more.

Just so someone else encountering this may have a clue.

(searching for SENDCOMMAND QUEUE to find some way of looking at the outbound queue was not )

 

cheers from the sanitation department.



[#] Mon Nov 20 2017 09:24:16 EST from IGnatius T Foobar @ Uncensored

Subject: Re: "Outbound Queue" invisible bounce sender to flood the database

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

Thanks for the very well written report. Would you be able to tell us how often the retry happened, and what the bounceback looked like? I'd like to ascertain whether the receiving server generated a 5XX error that should have told Citadel to give up, or a 4XX error that would have told Citadel to keep trying.

Older versions of Citadel Server would retry after 15 minutes, doubling the retry time after each attempt. Versions published starting this year (2017) simply retry every 30 minutes for the first four hours, and then every four hours for the next five days. But if it gets a 5XX undeliverable error from the receiving server, it's supposed to stop immediately.

[#] Mon Nov 20 2017 09:47:45 EST from Freakdog @ Dog Pound BBS II

Subject: Re: "Outbound Queue" invisible bounce sender to flood the database

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

* Log into the GUI as an admin (Aide/Sysop),
* Go to the *Administration* menu item on the left
* Select *Edit site-wide configuration*
* Go to the *Settings*B tab
* At the bottom, you'll see a check box next to *Automatically delete

committed database logs*
* Put a check mark in the box and click the button marked *Save changes*


That's it...

[#] Mon Nov 20 2017 14:35:53 EST from IGnatius T Foobar @ Uncensored

Subject: Re: "Outbound Queue" invisible bounce sender to flood the database

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

"Automatically delete committed database logs" is the default out-of-the-box configuration. I'm not sure why anyone would change it unless they were intentionally backing up the logs before deleting them.

[#] Mon Nov 20 2017 16:31:53 EST from kinetix @ Uncensored

Subject: Re: "Outbound Queue" invisible bounce sender to flood the database

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

I have been changing it on my installs precisely for the backup reason - it seems to be the only way to do it safely, according to the docs, without shutting citadel down?  Are there easier ways without causing the bit of system disruption?

 

Mon Nov 20 2017 02:35:53 PM EST from IGnatius T Foobar @ Uncensored Subject: Re: "Outbound Queue" invisible bounce sender to flood the database
"Automatically delete committed database logs" is the default out-of-the-box configuration. I'm not sure why anyone would change it unless they were intentionally backing up the logs before deleting them.

 



[#] Mon Nov 20 2017 17:14:29 EST from kinetix @ Uncensored

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

Hi all,

Just downgraded my citadel instances to the current stable tarballs from easyinstall, so citadel's at 911, webcit's at 904, libcitadel's at 904.

Removed the citadel networking config & room between my two systems, added in with a new password, a new test room, and am still seeing each system polling each other every 60 seconds whether they need to or not.

Any idea why the network run frequency isn't being honoured?

Thanks!

 

Wed Nov 08 2017 06:42:49 PM EST from kinetix @ Uncensored

Howdy all,

A couple of odd things - I'll first mention I'm running the git version of everything - one system was pulled on Oct. 25, the other on the 28th - I don't think these issues relate to any recent changes that I can see.

First - I had my home/test citadel drop port 504 unexpectedly.  I have no idea what could have changed it except that I restarted citadel via webcit.  Had to run setup again and setup saw that the port configured was 0.  

The other interesting thing going on is that I set up networking between the two systems, and regardless what I put in the "Network run frequency" setting, each system is polling the other every 60 seconds.  Each system is also keeping a 0 byte file in their network/spoolin directories.  I can't recall if that's normal behaviour or not.

The polling thing seems to be weird, but legitimate messages are moving between the systems.  

Is it worth posting logs relating to either of these?  The port 504 dropping thing doesn't seem to have any interesting log entry around it (citadel shut it down cleanly during it's restart and then never opened it again)... the network polling thing I figure might already be known?

 



 



[#] Mon Nov 20 2017 17:24:29 EST from IGnatius T Foobar @ Uncensored

Subject: Re: "Outbound Queue" invisible bounce sender to flood the database

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

I have been changing it on my installs precisely for the backup reason -
it
seems to be the only way to do it safely, according to the docs, without

shutting citadel down?  Are there easier ways without causing the bit of

system disruption?

According to the Berkeley DB docs, you are supposed to be able to back up the database while it is running, as long as you ensure the data files (cdb.*) are backed up *before* the log files (log.*)

That having been said ... that's not how I do it either. :) The virtual machine which runs my production Citadel instance (this one) has its virtual disks on a btrfs volume, which is snapshotted nightly and then the snapshots go off-host to a backup server. You could also make this work by running Citadel directly on a btrfs volume.

I see that Berkeley DB now has a "db_hotbackup" utility, which looks promising.
I haven't tried it.

Does anyone know if Linux/Unix has an API call that can tell a program the amount of free space left on a filesystem? Or does it have to be done by tediously mucking about with command line calls? It probably wouldn't be too hard to make Citadel Server refuse to accept incoming messages if there is no room to store them.

[#] Tue Nov 21 2017 03:02:33 EST from techacq1Aa @ Uncensored

Subject: Re: "Outbound Queue" invisible bounce sender to flood the database

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

 

Mon Nov 20 2017 09:24:16 EST from IGnatius T Foobar @ Uncensored Subject: Re: "Outbound Queue" invisible bounce sender to flood the database
Thanks for the very well written report. Would you be able to tell us how often the retry happened, and what the bounceback looked like? I'd like to ascertain whether the receiving server generated a 5XX error that should have told Citadel to give up, or a 4XX error that would have told Citadel to keep trying.

Older versions of Citadel Server would retry after 15 minutes, doubling the retry time after each attempt. Versions published starting this year (2017) simply retry every 30 minutes for the first four hours, and then every four hours for the next five days. But if it gets a 5XX undeliverable error from the receiving server, it's supposed to stop immediately.

I remember fondly the old double-time delay,
but this tempo was an unwavering  ~60 seconds.

 

In the users inbox, the reply has a sender (Citadel)

and a subject:

Subject: Delivery Status Notification (Failure)

A message you sent could not be delivered to some or all of its recipients
due to prolonged unavailability of its destination(s).
Giving up on the following addresses:

clarisse@abandoned.domain.com: 421 (Domain service not available, closing transmission channel)

 

She always cc: herself to ensure she gets a copy,

  and perhaps this was the crux of the issue - a cc: that trips the system into retrying the target when it sends the failure message, perhaps ?

just a thought

 

And to answer freakdog's suggestion: (thanks)

I never touched the autodelete logs - but checked again anyway to make sure that it is still set to delete and it is indeed set to delete.

 

I was first alarmed that it might be a hack,
but my email counter showed nothing going out

I have a script reporting the number of mails sent logged every few minutes on an admin webpage,
maybe I will add a  df  function to that log so I won't wake up to this again.
Mondays are bad enough without that kind of a hangover

My other concern was how I had set up the kvm with a detachable partition for /usr/local/citadel/data/
 but that was not a problem - it worked even after it hit the limits.

Now to automate the backup of this particular partition.....



[#] Tue Nov 21 2017 14:24:26 EST from kinetix @ Uncensored

Subject: Re: "Outbound Queue" invisible bounce sender to flood the database

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

Thanks for pointing that out!  I read the man page around it and thought I'd see what I got.  Unfortunately, the tool doesn't like something about the db itself:

root@home:/storage/citadel/data# db_hotbackup -v -d . -h . -b ../testbackup/        
db_hotbackup: BDB5032 hot backup started at Tue Nov 21 11:20:16 2017
db_hotbackup: BDB5129 Cannot copy data from a PRIVATE environment
db_hotbackup: BDB5043 HOT BACKUP FAILED!

I'm pretty confused by this, but know nothing about how 'environments' are set up in Berk DB.  I see references to it in citadel's database.c, but other than that, I'm at a loss as to what makes it not work here.

 

Mon Nov 20 2017 05:24:29 PM EST from IGnatius T Foobar @ Uncensored Subject: Re: "Outbound Queue" invisible bounce sender to flood the database
I see that Berkeley DB now has a "db_hotbackup" utility, which looks promising.
I haven't tried it.

 



[#] Thu Nov 23 2017 03:41:48 EST from marisag @ Uncensored

Subject: Directory rooms?

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

How do I set up file directory rooms? I enabled it in the room config but do=
n't see any upload/download buttons...

Marisa



[#] Thu Nov 23 2017 06:55:29 EST from techacq1Aa @ Uncensored

Subject: Re: Directory rooms?

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

clicking on the  word "files" in the upper header may send you to an upload/directory list,

as visible above just under where it says "Citadel Support"

if you gave the file directory a name, hopefully



[#] Thu Nov 23 2017 08:49:19 EST from marisag @ Uncensored

Subject: Re: Directory rooms?

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

Great, that took care of it. Thanks!



[#] Thu Nov 23 2017 08:50:36 EST from marisag @ Uncensored

Subject: Setting up citadel to answer for regular telnet connections?

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

It says to look in the install.txt file for instructions on doing this, but I installed via the easy install and don't have that file. Any pointers?



[#] Thu Nov 23 2017 10:57:39 EST from kinetix @ Uncensored

Subject: Re: Setting up citadel to answer for regular telnet connections?

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

You bet!

http://citadel.org/doku.php/documentation:system_administration_manual?s[]=telnet#bypassingtheloginprompt

 

Thu Nov 23 2017 08:50:36 AM EST from marisag @ Uncensored Subject: Setting up citadel to answer for regular telnet connections?

It says to look in the install.txt file for instructions on doing this, but I installed via the easy install and don't have that file. Any pointers?



 



[#] Thu Nov 23 2017 22:30:09 EST from techacq1Aa @ Uncensored

Subject: Re: "Outbound Queue" invisible bounce sender to flood the database

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

Again,
this time I was a bit better prepared.

Two entries in the outbound,

one original email to be resent
+ one error message about 421 failure (still to be sent !)

 

Both entries are utterly invisible in

..../dotskip?room=__CitadelSMTPspoolout__&view=11

and only visible in

..../dotskip?room=__CitadelSMTPspoolout__&view=1

not yet sure what that implies.

 

The message about ERROR 421 Domain service not available
compelled me to check again the email destination (using dig)
and I see no issues there, MX address seems okay, etc.

 

Ultimately there remain two things for me to solve:

1.
Have I have misconfigured some kind of disjoint
between citadel and the DNS server such that Citadel calls up
and the DNS simply fails to respond, tricking citadel into calling the destination unavailable ?

I will reserve some spare time to try to replicate this failure and examine it using an test email
 and not wait for another bounce event.

 

2.
I should rewire this 421 failure so that the system stops resending itself failed deliveries,
a to-do list thingy - time to read more on SENDCOMMAND.

 

The answer is out there and I will keep looking.

 

Tue Nov 21 2017 03:02:33 EST from techacq1Aa @ Uncensored Subject: Re: "Outbound Queue" invisible bounce sender to flood the database

 

In the users inbox, the reply has a sender (Citadel)

and a subject:

Subject: Delivery Status Notification (Failure)

A message you sent could not be delivered to some or all of its recipients
due to prolonged unavailability of its destination(s).
Giving up on the following addresses:

clarisse@abandoned.domain.com: 421 (Domain service not available, closing transmission channel)

 

She always cc: herself to ensure she gets a copy,

  and perhaps this was the crux of the issue - a cc: that trips the system into retrying the target when it sends the failure message, perhaps ?

just a thought

 

And to answer freakdog's suggestion: (thanks)

I never touched the autodelete logs - but checked again anyway to make sure that it is still set to delete and it is indeed set to delete.

 



[#] Fri Nov 24 2017 13:38:18 EST from richardb49 @ Uncensored

Subject: Installation woes

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

The server I had been running Citadel died two days ago. I put together another computer, installed Debian 9 and from root did the traditional apt-get install citadel-suite setting the IP address to my server's IP, set the admin user name and password.  The one thing I did not do was first create a separate user for Citadel.

Brought up Webcit, entered the user name and password I put in for the admin, got a "Please wait" and several beeps from the speaker. Nothing else. Went to restart the computer, browsed to Webcit, tried to log in and get a wrong password. I have purged and reinstalled, I have wiped the hard drive and started over several times over the last two days. Sometimes it says the citadel server is not running and I manually restart it. All to no avail. I love the program, I just am at my wits end on what should have been a simple install.

Checked the system log and it has:

[ 3710.658289] citserver[1758]: segfault at 0 ip b6ffb5a6 sp b41f4e94 error 4 in libc-2.24.so[b6ebd000+1b1000]
[ 3711.174977] citserver[1778]: segfault at 0 ip b6ffb5a6 sp bfd096f4 error 4 in libc-2.24.so[b6ebd000+1b1000]
[ 3711.633529] citserver[1781]: segfault at 0 ip b6ffb5a6 sp bfd096f4 error 4 in libc-2.24.so[b6ebd000+1b1000]
[ 3789.535293] TCP: request_sock_TCP: Possible SYN flooding on port 25. Sending cookies.  Check SNMP counters.
[ 4315.358799] citserver[1787]: segfault at 0 ip b6ffb5a6 sp b42fd044 error 4 in libc-2.24.so[b6ebd000+1b1000]
[ 4319.219975] citserver[1819]: segfault at 0 ip b6ffb5a6 sp b41fc044 error 4 in libc-2.24.so[b6ebd000+1b1000]
[ 4321.931053] citserver[1823]: segfault at 0 ip b6ffb5a6 sp b42fd044 error 4 in libc-2.24.so[b6ebd000+1b1000]
[ 5073.461855] TCP: request_sock_TCP: Possible SYN flooding on port 25. Sending cookies.  Check SNMP counters.

Any help would be appreciated.



[#] Fri Nov 24 2017 16:08:46 EST from techacq1Aa @ Uncensored

Subject: Re: Installation woes

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

Hi Richard,

Me too - Debian 9 has issues, the packaged debs are hopeless.

If you use the easy-install everything may fall more or less into place.

Just be sure you do install all the prerequisites. I know, I have been slogging into Debian 9 also.

Right now I am degraded to a SolydX 8 (=debian8) due to the problems getting past the post on Debian 9.1 and its SolydX derivative.

Citadel does install a lot easier in Ubuntu, (I am very unhappy with Ubuntu, but easy install on Xubuntu seems to work properly.)

Fri Nov 24 2017 13:38:18 EST from richardb49 @ Uncensored Subject: Installation woes

The server I had been running Citadel died two days ago. I put together another computer, installed Debian 9 and from root did the traditional apt-get install citadel-suite setting the IP address to my server's IP, set the admin user name and password.  The one thing I did not do was first create a separate user for Citadel.

Brought up Webcit, entered the user name and password I put in for the admin, got a "Please wait" and several beeps from the speaker. Nothing else. Went to restart the computer, browsed to Webcit, tried to log in and get a wrong password. I have purged and reinstalled, I have wiped the hard drive and started over several times over the last two days. Sometimes it says the citadel server is not running and I manually restart it. All to no avail. I love the program, I just am at my wits end on what should have been a simple install.

Checked the system log and it has:

[ 3710.658289] citserver[1758]: segfault at 0 ip b6ffb5a6 sp b41f4e94 error 4 in libc-2.24.so[b6ebd000+1b1000]
[ 3711.174977] citserver[1778]: segfault at 0 ip b6ffb5a6 sp bfd096f4 error 4 in libc-2.24.so[b6ebd000+1b1000]
[ 3711.633529] citserver[1781]: segfault at 0 ip b6ffb5a6 sp bfd096f4 error 4 in libc-2.24.so[b6ebd000+1b1000]
[ 3789.535293] TCP: request_sock_TCP: Possible SYN flooding on port 25. Sending cookies.  Check SNMP counters.
[ 4315.358799] citserver[1787]: segfault at 0 ip b6ffb5a6 sp b42fd044 error 4 in libc-2.24.so[b6ebd000+1b1000]
[ 4319.219975] citserver[1819]: segfault at 0 ip b6ffb5a6 sp b41fc044 error 4 in libc-2.24.so[b6ebd000+1b1000]
[ 4321.931053] citserver[1823]: segfault at 0 ip b6ffb5a6 sp b42fd044 error 4 in libc-2.24.so[b6ebd000+1b1000]
[ 5073.461855] TCP: request_sock_TCP: Possible SYN flooding on port 25. Sending cookies.  Check SNMP counters.

Any help would be appreciated.



 



[#] Fri Nov 24 2017 21:05:47 EST from richardb49 @ Uncensored

Subject: Re: Installation woes

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

Thanks for the reply. I will try Debian 8.9 in the morning and post my results. I even bought a new PC in desperation this afternoon in case it was hardware related. I did not realize how many things you have to disable in an HP machine to get it to take anything but Windoze 10. Debian 9 still has issues on the new computer: hangs up, hard to set ethernet ports (no more eth0, eth1, etc.).

Once I get it up and running and all my users accounts are set up again, my next quest will be to figure if it is possible to retrieve the Citadel database on the old hard drive (I have been able to access it by connecting it to a working computer).

 



Go to page: First ... 29 30 31 32 [33] 34 35