Ok...so we've been discussing 2 issues...one is the issue of networking getting hung up, another (this was discussed in the Citadel Support room) is the occasional database related shutdown.

Mine shutdown, yesterday...I'm attaching the log from just yesterday (haven't perused it, yet, but the only reason mine usually shuts down is DB related).

Guess I'll have to dig into git, a little, and have it grab the master tree.

fd, i think its ok to handle this over here; most users in cit sup don't care about this topic, and it spoils their backlog.

I did some stabilization and test work on the git master source last week;

It has (again ;-) some major overhaul of the ignet stuff - this time the config & queueing part.

I'd like to know whether it fixes your incident.

We also now have some more debugging for the NTT-List which may be the source of your problems.

Plus some rework of the code that manages the seen-lists - reducing frequency and hopefully the database assertions.

Here's my citadel log for the last week.

Sat Mar 23 2013 05:37:58 AM EDT from dothebart @ Uncensored

 

Yes. Now, if I could only figure out why I wind up with a ton of spoolfiles that periodically build up in my spoolout directory until I restart Citadel, I'd be happy.

The one thing that continually seems to bite me in the arse.

I think I may have to just schedule periodic citadel restarts.

well, solving it should be the better alternative.

turn on debug logging for these facilities, and find out whats going on.

do emails & rss continue working?

SMTP based, i believe so (I don't do as much via SMTP), RSS, yes. It's just the Citadel Networking that seems borked.

I've got logging enabled...I can send the log, but I'm not sure what date/time things stopped working.

Fri Mar 22 2013 14:36:51 EDT from Freakdog @ Dog Pound BBS II

 

Tue Mar 19 2013 08:57:18 AM EDT from dothebart @ Uncensored

so now you can send messages back & forth, and they don't go via smtp?

Yes. Now, if I could only figure out why I wind up with a ton of spoolfiles that periodically build up in my spoolout directory until I restart Citadel, I'd be happy.

The one thing that continually seems to bite me in the arse.

I think I may have to just schedule periodic citadel restarts.

well, solving it should be the better alternative.

turn on debug logging for these facilities, and find out whats going on.

do emails & rss continue working?

Tue Mar 19 2013 08:57:18 AM EDT from dothebart @ Uncensored

so now you can send messages back & forth, and they don't go via smtp?

Yes. Now, if I could only figure out why I wind up with a ton of spoolfiles that periodically build up in my spoolout directory until I restart Citadel, I'd be happy.

The one thing that continually seems to bite me in the arse.

I think I may have to just schedule periodic citadel restarts.

so now you can send messages back & forth, and they don't go via smtp?

good question - i've read they need to share the GAB? Though I didn't find anything doing this automatic - maybe you need to pair them manualy? i guess IG should know more.

Sun Dec 16 2012 11:13:44 PM EST from IGnatius T Foobar @ Uncensored

i see messages

Yes...things seem to be working again.

Wed Jan 09 2013 09:43:44 AM EST from dothebart @ Uncensored

 

Mon Jan 07 2013 11:09:48 EST from Freakdog @ Dog Pound BBS II

 

Sun Jan 06 2013 12:00:57 PM EST from dothebart @ Uncensored

-x once took a loglevel...

it now accepts the very same list as the environment variable.

note that you need to export it.

D'oh! Didn't think to "export" since the CITADEL_DIR variable wasn't, though it seemed to be pulled in.

So, if I have "networktalkingto,networkqueue,networkclient,4 set for CITADEL_LOGDEBUG or -x, will it still use log level 4 in addition to the 3 network monitoring settings? Just wondering.

no, loglevel configuration isn't supported anymore.

however, you can still choose which logverbosity you want in your syslog config.

the new facility ommits all debug-logging by default, you can re-enable this on a per application base with the above mechanism.

else citserver would produce a lot of load on syslog, which syslog then needs to discard according to his loglevel, so not sending it if its unwanted saves resources...

Someone ought to update the documentation, then. The Citadel Administration Manual still shows:

The next step is to arrange for the server to start. The “citserver” program is the main Citadel server. Before we cover the recommended method of starting the server, let's examine its usage options:

citserver [-hHomeDir] [-xDebugLevel] [-tTraceFile] [-lLogFacility] [-d] [-f]

And

”-xDebugLevel” - Set the verbosity of trace messages printed. When -x is used, it will suppress messages sent to syslog (see below). In other words, syslog will never see certain messages if -x is used. Normally you should configure logging through syslog, but -x may still be useful in some circumstances. The available debugging levels are:

• 0 - Emergency condition; Citadel will exit immediately
• 1 - Severe errors; Citadel may be unable to continue without attention
• 2 - Critical errors; Citadel will continue with degraded functionality
• 3 - Error conditions; Citadel will recover and continue normally
• 4 - Warning messages; Citadel will continue normally
• 5 - Normal operational messages
• 6 - Informational messages, progress reports, etc.
• 7 - Debugging messages; extremely verbose

under the "Configuring your host system to start the service" section of:

http://www.citadel.org/doku.php/documentation:system_administration_manual#citadel.admininistration.manual

This may lead to some confusion (it certainly did for me), otherwise.

Mon Jan 07 2013 11:09:48 EST from Freakdog @ Dog Pound BBS II

 

Sun Jan 06 2013 12:00:57 PM EST from dothebart @ Uncensored

-x once took a loglevel...

it now accepts the very same list as the environment variable.

note that you need to export it.

D'oh! Didn't think to "export" since the CITADEL_DIR variable wasn't, though it seemed to be pulled in.

So, if I have "networktalkingto,networkqueue,networkclient,4 set for CITADEL_LOGDEBUG or -x, will it still use log level 4 in addition to the 3 network monitoring settings? Just wondering.

no, loglevel configuration isn't supported anymore.

however, you can still choose which logverbosity you want in your syslog config.

the new facility ommits all debug-logging by default, you can re-enable this on a per application base with the above mechanism.

else citserver would produce a lot of load on syslog, which syslog then needs to discard according to his loglevel, so not sending it if its unwanted saves resources...

Sun Jan 06 2013 12:00:57 PM EST from dothebart @ Uncensored

-x once took a loglevel...

it now accepts the very same list as the environment variable.

note that you need to export it.

D'oh! Didn't think to "export" since the CITADEL_DIR variable wasn't, though it seemed to be pulled in.

So, if I have "networktalkingto,networkqueue,networkclient,4 set for CITADEL_LOGDEBUG or -x, will it still use log level 4 in addition to the 3 network monitoring settings? Just wondering.

-x once took a loglevel...

it now accepts the very same list as the environment variable.

note that you need to export it.

Ok, I have both of these in my init script, and maybe they're conflicting:

CITADEL_LOGDEBUG=networktalkingto,networkqueue,networkclient

and

CITADEL_DIR/citserver -d -h$CITADEL_DIR -x4 -llocal6

I do have:

CITADEL_DIR/citserver -d -h$CITADEL_DIR -x6 -llocal6

commented out at the moment.

you can set it using either environment or commandline parameters, in your init script, whatever you prefer.

its

CITADEL_LOGDEBUG

or

-x

and then, coma separated the list

as

sendcommand LOGP

gives you.

you should however see the active sessions on your side using

sendcommand RWHO

Sun Dec 16 2012 11:13:44 PM EST from IGnatius T Foobar @ Uncensored

i see messages

Yes...things seem to be working again.

Sat Dec 01 2012 04:35:55 AM EST from dothebart @ Uncensored

 

Wed Nov 28 2012 13:51:18 EST from Freakdog @ Dog Pound BBS II

Haven't seen anything from Uncensored in a few days...is something amiss?

ah, there it is.

in general, can you please turn on the citadel networkers debug logging in webcit?

I've enabled all three networking options. Should this happen again, we can look in the logs and see what's up.

Sat Dec 01 2012 13:15:58 EST from Freakdog @ Dog Pound BBS II

 

Sat Dec 01 2012 04:35:55 AM EST from dothebart @ Uncensored

 

Wed Nov 28 2012 13:51:18 EST from Freakdog @ Dog Pound BBS II

Haven't seen anything from Uncensored in a few days...is something amiss?

ah, there it is.

in general, can you please turn on the citadel networkers debug logging in webcit?

I've enabled all three networking options. Should this happen again, we can look in the logs and see what's up.

please don't forget the NetworkTalkingTo logging...

Sat Dec 01 2012 04:35:55 AM EST from dothebart @ Uncensored

 

Wed Nov 28 2012 13:51:18 EST from Freakdog @ Dog Pound BBS II

Haven't seen anything from Uncensored in a few days...is something amiss?

ah, there it is.

in general, can you please turn on the citadel networkers debug logging in webcit?

I've enabled all three networking options. Should this happen again, we can look in the logs and see what's up.

Wed Nov 28 2012 13:51:18 EST from Freakdog @ Dog Pound BBS II

Haven't seen anything from Uncensored in a few days...is something amiss?

ah, there it is.

in general, can you please turn on the citadel networkers debug logging in webcit?

Thu Nov 29 2012 09:00:24 EST from Freakdog @ Dog Pound BBS II

Cleared out some files in the spoolout directory here...let's see if this makes it across.

Looks like something got corrected, here.

Fri Nov 30 2012 08:04:18 EST from Freakdog @ Uncensored

 8.16 via easyinstall, and yes, quite frequently, there is a SYS_CitNetworker process. Not always for Uncensored, though.

Correction...8.15 via easyinstall. ;-)

Thu Nov 29 2012 16:38:05 EST from dothebart @ Uncensored

is your ip?...

An unknown Citadel server called "" attempted to connect from 206.253.164.12 [206.253.164.12

Nope. I'm coming from 108.245.117.203.

Thu Nov 29 2012 16:25:54 EST from dothebart @ Uncensored

 

Thu Nov 29 2012 14:49:14 EST from Freakdog @ Uncensored

I do, indeed...nothing's come into any rooms I share with Uncensored since November 26.

-rw-r--r-- 1 libuuid libuuid  53374 Nov 29 16:11 comalies
-rw-r--r-- 1 libuuid libuuid  83520 Nov 29 16:11 dogpound2
-rw-r--r-- 1 libuuid libuuid      0 Nov 29 16:12 outgesourced

hm, indeed, seems as if you don't retrieve your package ;)

you're running 8.1x meanwhile right?

if you're an aide user... in your online users, is there a citadel networker all the time?

And either Uncensored isn't picking up its packages, or we're not sending them.

8.16 via easyinstall, and yes, quite frequently, there is a SYS_CitNetworker process. Not always for Uncensored, though.

is your ip?...

An unknown Citadel server called "" attempted to connect from 206.253.164.12 [206.253.164.12

Thu Nov 29 2012 14:49:14 EST from Freakdog @ Uncensored

I do, indeed...nothing's come into any rooms I share with Uncensored since November 26.

-rw-r--r-- 1 libuuid libuuid  53374 Nov 29 16:11 comalies
-rw-r--r-- 1 libuuid libuuid  83520 Nov 29 16:11 dogpound2
-rw-r--r-- 1 libuuid libuuid      0 Nov 29 16:12 outgesourced

hm, indeed, seems as if you don't retrieve your package ;)

you're running 8.1x meanwhile right?

if you're an aide user... in your online users, is there a citadel networker all the time?

hm, networking to outgesourced is working fine.

afair you share citadel support?

Sorry folks, "enjoythesilence" was me. A little trouble migrating the install over to a new server, but it should be good now. If it is still showing up in your logs let me know. 

No worries, Matt...it did clear up...just wasn't sure, at the time, what was up.

Mon Oct 15 2012 13:23:39 EDT from Freakdog @ Dog Pound BBS II

 
In the meantime, we're online and networking again.

I'm seeing what appear to be successful connects from your system to mine (trada) but I'm not seeing anything from my systems making it here to Uncensored...

Jame

interesting usecase.

are you running latest stable on all systems so far?

and FD, if I got that correctly, you're upgrading with your merger to the new box?

Thu Oct 04 2012 10:35:03 EDT from Freakdog @ Uncensored

Well, they suspended my account and services some time yesterday...time for me to build out that other server...post haste. Feh.

It looks like this was a bad time for me to do a post on my local system in order to test if the networking issue between my systems has been resolved...<g> The linkage is my main local node (rocasa) to my public server (trada) to dogpound2:  dogpound2 <-> trada <-> rocasa

By way of testing that (since it appeared to have been working), I posted a message about an update to Net::Citadel on my rocasa node (aka mail.rocasa.us) in Citadel Support.  That did make it to trada but not further;  investigating that is when I found that dogpound2 is currently down....

I'll check here periodically to see when you're able to get your system back on line.

Jame

No worries, Matt...it did clear up...just wasn't sure, at the time, what was up.

Sorry folks, "enjoythesilence" was me. A little trouble migrating the install over to a new server, but it should be good now. If it is still showing up in your logs let me know. 

LOL...nice.

Who is the sysop of "enjoythesilence"? I've been getting connections from the node, but have not configured networking with any nodes with that name.

Please contact me to arrange networking, if you wish to connect with/through Dog Pound II

hm, which kernel version does 5.8 have? 2.6.18? did you write about your problems?

Ok...chances are that it will happen as of the 28th...I expect that the cloud instance on which DP2 currently runs will be shut off and destroyed some time on that date.

I'll be building a Centos 6 system onto which I'll migrate DP2 once I've returned from Texas to Indiana, but because my server in Indiana is running CentOS 5.8 and I was completely unable to get the latest Citadel versions running under that version, I'll need to build out Centos 6.3 to make it happen...but I'll do what I can to minimize the amount of time DP2 is down, in order to get networking back into play as quickly as possible.

if you' re an aide, system threads are shown.

have a look at the smtp queue, some goodies there ;-)

Ok...I'll back 'er up and upgrade her overnight while I'm working.

networking should work fine with 8.14.

you can however try whether splice works the right way on your site by adding

export LIBCITADEL_ENABLE_SPLICE=1

you should check whether the re-syncing and collecting the files into the spool works properly.

Sat Oct 01 2011 07:15:46 AM PDT from matt @ comalies

@Freakdog: port 504 on dogpound2.citadel.org doesn't appear to be reachable, even though I can see webcit on it. 

Should be good now. I migrated my citadel from my home server to a cloud instance so that I could work on upgrading it to a newer version.

ok, new citadel networking code & backend code is deploied on uncensored.

in general news, please upgrade to the latest source and revalidate:

  - on recent linuxes you can use splice enabled code; which can be enabled by adding this to your environment: LIBCITADEL_ENABLE_SPLICE=1

  - you can revalidate this by pressing the new circle icon in the room preferences sharing tab; the spool directory should then fill up with message files, which later get combined to one file.

if the one file thing works, splice is ok to use for you.

since we had some troubles in the networking code, i've re-sent the messages in citadel support and ignet unlimited to your nodes (using that very same button ;-)

-- 
o__
_> / _
(_) \(_)...

fd,

current git contains some fixups for citadel networking.

there also is a new <resync> button in webcit, which you might want to use to get stuff resync'ed...

-- 
o__
_> / _
(_) \(_)...

Thu Jul 05 2012 05:41:33 AM EDT from Freakdog @ Uncensored

Well, something's up, that's for sure. I'm running 8.11 on the test system, and 8.05 on my main system...and both appear to be working just fine.

Ok...this came through. Interesting.

yep, there were, two of them.

we probably still have troubles with building the package from the bunch of message spool files.

well, maybe you should run easy install once, to get a fresh copy of curl in /usr/local/ctdlsupport that is up to date...  and then compile your tarball with added  parameters

are you running rpms? or easy install? since c-ares doesn't have many dependencies... well except maybe those which openbsd mixes up...

Unfortunately, I can't. c-ares on CentOS 5.8 doesn't cover MX lookups, and my efforts to install something newer still didn't work for me, somehow.

please try upgrading to latest git, we had some issues about file descriptor leaks, and concurency checks which will make stuff get over slow.

yes, got here ;-)

And we're back.

Wed Apr 18 2012 07:31:00 PM GMT from IGnatius T Foobar @ Uncensored

I noticed in my logs that comalies is now networking with uncnsrd via IPv6!
*< kermit flails >*

:)

IPv6 has been on since last year. 

Why not "trek wars"?  "star trars"?  "Star warks"? "Kirk v. Skywalker"?

I said "an up and coming competitor"... ;-)

You guys know that I wouldn't go to work for M$...besides...what would they do with an AIX/Linux admin, anyhow?

I'm working for HP Cloud Services.

And so it begins.

well, that possible overrun is the number of sessions in total (i.e. SMTP counts here too)

since it seems to work most of the time, and sometimes not, I thought it might occur on your side...

so, the new networker code is up and running.

for some reasons there were some contact attempts where the networker bailed out with 'dp2: so the : is to much.

imho it just can have been added by the server... maybe a connection overload?

Fri Jan 20 2012 03:14:51 PM EST from IGnatius T Foobar @ Uncensored

Just making sure that comms are still working.

Loud and clear on this end.

WOOT!!!

Yup:

; <<>> DiG 9.6-ESV-R4-P3 <<>> dogpound2.citadel.org +multiline +nocomments +nocmd +noquestion +nostats +search
;; global options: +cmd
dogpound2.citadel.org.    60 IN CNAME bbs.bubbanfriends.org.
bbs.bubbanfriends.org.    289 IN A 69.212.163.244

Tue Nov 22 2011 10:11:58 PM EST from IGnatius T Foobar @ Uncensored

dothebart: Is that fix in 8.01?

It's definitely in 8.02, and I just posted tarballs of that release tonight.

Excellent...Going to download and install it, now.

its definitely in git-master; dunno whether its in 8.01.

a 8.02 might want to wait for us to find the imap troubles in conjunction with the php imap client implementation...

Tue Nov 01 2011 16:49:37 EDT from Freakdog @ Dog Pound BBS II

dothebart: Is that fix in 8.01?

Freakdog,

   And/or anyone else who sees this message...

   I'm connected to the Citadel rooms via Freakdogs system.   The rooms appear to be connected but if they're getting updated messages then they're on some kind of delay because I'm getting old messages as new...

   I maintain a connectiion the Citadel Support mailing list at my Gmail account & that has been seeing much newer messages, by several days, than what are actually showing up here...

Jame

Sun Oct 16 2011 09:53:53 EDT from Freakdog @ Dog Pound BBS II Subject: Re:

Where is that list kept, how is it kept, and why is it that when I erased what appeared to be a spool file from the spoolout directory, networking with the node in question started working?

the list is kept in memory; its set by either the network queuing service itself, an inbound or an outbound networking connection.

once that connection is closed, the lock should be released.

However, I found a possible race very recently which I fixed by removing global vars.

unless you enable coring, and kill a stuck citserver with SIG11 and share some delights with us, we probably won't shed more light on why removing the file unfucked it.

Thu Oct 13 2011 15:25:36 EDT from Freakdog @ Dog Pound BBS II Subject: Re:

Do we have lock files for networking sessions?

we keep a list of nodes that we are currently talking to / are talking to us.

so, its sort of a lock.

Freakdog: This is what happens when comalies tries to talk to dogpound2. 

Oct 11 01:48:08 localhost citserver[7357]: >200 dogpound2 Citadel server ready.

Oct 11 01:48:08 localhost citserver[7357]: <NETP comalies|XXXXXX

Oct 11 01:48:08 localhost citserver[7357]: >552 Already talking to comalies right now

Stale lockfile maybe?

Got it squared away...lost power last week, and there's an issue with bringing the BBS IP back up before the Cit software.

That issue should now be resolved.

Sat Oct 01 2011 10:15:46 AM EDT from matt @ comalies

@Freakdog: port 504 on dogpound2.citadel.org doesn't appear to be reachable, even though I can see webcit on it.

@Freakdog: port 504 on dogpound2.citadel.org doesn't appear to be reachable, even though I can see webcit on it. 

hm, I probably broke networking for master

fixed with 2506fdb3c8df79cd2055bb3ca8306494d6d61e94

messages woludn't be netqueued...

Freakdog,

Have not received a reply to my email to you about not receiving network transfers apparantly not working;  no new messages from your system and new messages from my (cobweb) system not showing up at Uncensored.   Didn't get a bounce on the email.   Found my login info here, and have found that your system apparantly has not received any of the new messages from Uncensored;  in Citadel Support, for instance.

Freakdog,

I'm not getting any new messages in any of the shared rooms from your your system (they do work locally;  but none of the new postings on my systems have made it to yours).  I'm not seeing anything in the logs about why, but I'm not entirely sure about where to look...  )

Should  I at least be seeing connection attempts from your system?  Although I'm not sure what you might have used for a domain name for that...

Jame

cobweb.citadel.org

cobweb.rocasa.us

Sun Jul 31 2011 16:11:05 EDT from Freakdog @ Dog Pound BBS II Subject: Re: New system connection

I've replied both here on my system and via email to the addresses you've supplied. :-)

Looks like it's working...  Thanks!

jame

Tue Aug 02 2011 21:56:54 EDT from IGnatius T Foobar @ Uncensored Subject: Re: New system connection

Welcome to the network, jame!

Thanks!

Sun Jul 31 2011 16:11:05 EDT from Freakdog @ Dog Pound BBS II Subject: Re: New system connection

I've replied both here on my system and via email to the addresses you've supplied. :-)

And I think I have it set up correctly now...<g>

 

 

Jame

 

I've replied both here on my system and via email to the addresses you've supplied. :-)

Fri Jul 22 2011 01:12:19 PM EDT from Freakdog @ Dog Pound BBS II Subject: Re: New system connection

I can...log into dogpound2.citadel.org, let me know what rooms you'd like to share, and we'll get things going.

I sent you a message about the initial set of rooms;  please let me know if you need more information...

Jame

Hi All!

I've set up a new Citadel system, node name cobweb, and would like to get connected into the Citadel network.   Is there someone who can help me with this?

Jame

sysop@rocasa.us

No idea, but I do get quite a few of these messages in Aide from time to time, like on May 18:

Connected to node "dogpound2:" but I was expecting to connect to node "dogpound2".

Hi Freakdog,

Do Mai 05 2011 09:42:28 EDT von Freakdog @ Dog Pound BBS II

Just need to know...from where would I obtain the RPMs? (I think I missed that somewhere, and didn't notice it on the citadel.org download page yet)

Navigate to http://download.opensuse.org/repositories/home:/homueller:/citadel/ and choose your distribution and architecture.

Best regards

Stefan

No idea, but I do get quite a few of these messages in Aide from time to time, like on May 18:

Connected to node "dogpound2:" but I was expecting to connect to node "dogpound2".

MyCastlerealm deomain is expireing on the 25th and will not be renewed. 

Freakdog,

   Go ahead and pull networking any time after Friday.

Ian,

  I've dumped a LOT of domains over the past 6 months.  I need to drop you an e-mail w/ an updated list.. My Castle realm and wall will both be dead for good as of Friday.

Sun Dec 19 2010 11:49:36 PM EST from IGnatius T Foobar @ Uncensored Subject: Networking is fixed

Networking is fixed.

Either updgrade to the latest Easy Install, or apply the patch at http://easyinstall.citadel.org/netfix.patch

Patched tarballs will be available soon.

W00T!!!  Fix applied and in use at DPII.

Is that it?  Maybe if I put my sleep statement in between the stop and the IP down.

I can remove it, though...and see how that works out.

Mon Jul 26 2010 04:15:39 AM EDT from saltine @ Uncensored Subject: peering?

Anyone interested in peering with me? unixwire.com

I've got some bandwidth.

Log into dogpound2.citadel.org, and let me know which rooms you'd like to share.

Does anyone know what happened to Heretic Flambe?  It appears that I've not been able to connect to it for nearly a year, so I'm going to pull the plug on its network connection, unless someone has a line to the owner and can get hold of them.

Sa Jun 19 2010 22:27:30 CEST von IGnatius T Foobar @ Uncensored

Over the last couple of months I've been brainstorming some ideas on how we can make the network architecture more dynamic, allowing nodes to join and resign at any time without any manual configuration required.

What about http://en.wikipedia.org/wiki/Zeroconf ?

Sat May 07 2011 04:52:02 AM EDT from dothebart @ Uncensored Subject: Re:

could you ask that stefan and ti_eugene in citadel support?

I can...thanks.

could you ask that stefan and ti_eugene in citadel support?

Do Mai 05 2011 09:42:28 EDT von Freakdog @ Dog Pound BBS II

Just need to know...from where would I obtain the RPMs? (I think I missed that somewhere, and didn't notice it on the citadel.org download page yet)

Срд Апр 20 2011 06:55:51 EDT от dothebart @ Uncensored Тема: Re: Citadel @ RH-based distros

can you please provide the URLs from the OBS-Rpms for those distributions?

 

Now I'm repackaging it using 7.85 sources

http://download.opensuse.org/repositories/home:/homueller:/citadel/

 

Fr Apr 22 2011 20:38:47 EDT von Freakdog @ Dog Pound BBS II Betreff: Re:

 

Fri Apr 22 2011 06:33:46 PM EDT from dothebart @ Uncensored Subject: Re:

Hey FD, you're running fedora, right?

would you mind testing those SuSE OBS RPMs?

Old server's running Fedora 5.

New server's running CentOS 5.6.  I can give 'em a try there, since I don't have anything live running there, yet.

that would be realy great, please give eugene feedback in citadel support about your findings.

TIA.

Fri Apr 22 2011 06:33:46 PM EDT from dothebart @ Uncensored Subject: Re:

Hey FD, you're running fedora, right?

would you mind testing those SuSE OBS RPMs?

Old server's running Fedora 5.

New server's running CentOS 5.6.  I can give 'em a try there, since I don't have anything live running there, yet.

Hey FD, you're running fedora, right?

would you mind testing those SuSE OBS RPMs?

Is that it?  Maybe if I put my sleep statement in between the stop and the IP down.

I can remove it, though...and see how that works out.

/var/log/citadel.log

Thu Feb 03 2011 08:40:38 AM EST from IGnatius T Foobar @ Uncensored

Ok, I restarted Citadel here ... any change?

Nothing new entered DP2.

*whalks away whistling*

Sun Dec 19 2010 11:49:36 PM EST from IGnatius T Foobar @ Uncensored Subject: Networking is fixed

Networking is fixed.

Either updgrade to the latest Easy Install, or apply the patch at http://easyinstall.citadel.org/netfix.patch

Patched tarballs will be available soon.

W00T!!!  Fix applied and in use at DPII.

Ok, I've got a clearer picture of what's happening now, and a fix should be forthcoming.  THe move to buffered I/O completely broke certain modes of transport that are used between servers.

Thu Nov 25 2010 09:54:01 AM EST from IGnatius T Foobar @ Uncensored

At the moment I'm trying to figure out why networking is all jammed up.

Still jammed?

I'm seeing this in my logs:

2010/12/08 11:20:27.511984 [LWP:1125] [:4] network: processing inbound queue
2010/12/08 11:20:27.512041 [LWP:1125] [:4] network: processing 14434 bytes from /usr/local/citadel/network/spoolin//uncnsrd.4cfecf435e61fc87
2010/12/08 11:20:27.512528 [LWP:1125] [:4] network_process_buffer() processing 1 bytes
2010/12/08 11:20:27.512538 [LWP:1125] [:4] Corrupt message ignored.  Length=1, firstbyte = 255, lastbyte = 255
2010/12/08 11:20:27.512550 [LWP:1125] [:4] network_process_buffer() processing 1 bytes
2010/12/08 11:20:27.512558 [LWP:1125] [:4] Corrupt message ignored.  Length=1, firstbyte = 255, lastbyte = 255
2010/12/08 11:20:27.512569 [LWP:1125] [:4] network_process_buffer() processing 1 bytes
2010/12/08 11:20:27.512577 [LWP:1125] [:4] Corrupt message ignored.  Length=1, firstbyte = 255, lastbyte = 255
2010/12/08 11:20:27.513198 [LWP:1125] [:4] network_process_buffer() processing 8175 bytes
2010/12/08 11:20:27.514751 [LWP:1125] [:4] CtdlSubmitMsg() called
2010/12/08 11:20:27.514764 [LWP:1125] [:4] ERROR: attempt to save message with NULL body
2010/12/08 11:20:27.514829 [LWP:1125] [:4] network: queue run completed

after ugrading to latest stable?

probably a problem with the change to the way netconfigs are handled now?

i've changed it not to use tmpfile which would write configs to /tmp/ but rather create the file in the target directory, and rename it.

by that way we don't loose netconfigs (rss feed urls pop collectors, mailing list subscribers, node links..) if the filesystem containing the netconfig directory is full.

maybe for some reason in non-.deb mode the paths aren't built correctly?

IG, maybe you should upgrade uncensored to latest stable too? /me just crashed it by some old bug :-(

Mon Jul 26 2010 04:15:39 AM EDT from saltine @ Uncensored Subject: peering?

Anyone interested in peering with me? unixwire.com

I've got some bandwidth.

Log into dogpound2.citadel.org, and let me know which rooms you'd like to share.

Anyone interested in peering with me? unixwire.com

MyCastlerealm deomain is expireing on the 25th and will not be renewed. 

Freakdog,

   Go ahead and pull networking any time after Friday.

Ian,

  I've dumped a LOT of domains over the past 6 months.  I need to drop you an e-mail w/ an updated list.. My Castle realm and wall will both be dead for good as of Friday.

I emailed you quite a while ago that I was pulling it - probably go lost somewhere in the bit bucket.

Does anyone know what happened to Heretic Flambe?  It appears that I've not been able to connect to it for nearly a year, so I'm going to pull the plug on its network connection, unless someone has a line to the owner and can get hold of them.