Topic: SME 6.0 / 7.0 rc2 Fetching Mail socket error

[abarron]

Hi

First I am a noob so please forgive / advise any transgressions. I haven't posted it as a bug because I don't think it is yet.

I have been using 6.0 for a couple of years as a home server for web and email and have been very happy with it.   I send email directly but collect email from my domain hosts remote POP3 mail box.  Everything worked fine until this week when mail was only being collected intermittently, I noticed the following in the mail log.

May  7 16:05:05 server fetchmail[3676]: socket error while fetching from mail.*.*
May  7 16:05:05 server fetchmail[3676]: Query status=2 (SOCKET)

This is actually since I updated to v7rc2 but the messages from v6 were similar.

I can collect the mail from the remote pop3 using my mail client so I don't think it is them and I can browse and download from my web site externally so I don't think it is the hardware.  In desperation I upgraded to v7rc2 which made no difference.  I can think of nothing that has changed (situation same before upgrade) to cause this.

Any help gratefully received.  BTW I have read http://forums.contribs.org/index.php?topic=28698.0
but I don't think it is the same problem although I don't completely understand that post.

Thanks

Andy

[abarron]

Just realised I didn't make it clear that every fetchmail attempt ends up with those errors although despite this a very small number appear to succeed.

I have done some reading and some experiments today and have discovered that if I invoke startmail manually it appears to work.  Now I am really confused because that AIUI is exactly what Cron does so what is the difference.

I produced a log having specified -v -v in fetchmail which I can post if anyone thinks that is helpful.

Please help.

Andy

[abarron]

In the vain hope that someone may be willing to help a noob here is another question.

When I invoke startmail from the command line logged in as root, I don't get any entries in the mail log, is this because it has worked or should I still be getting an entry.

Is it that the log gets put somewhere else because I am logged in as root and if so where?  I do get an entry in the message log indicating that a session was opened and closed for qmailr so I am guessing I don't get a message in the mail log on success but confirmation would be nice  )

Andy

[CharlieBrady]

Please help.

Every time you post a message you are told:

  Please post bugs and potential bugs in the bug tracker - Thank You.

You would help yourself if you paid attention to that advice.

[abarron]

Every time you post a message you are told:

  Please post bugs and potential bugs in the bug tracker - Thank You.

You would help yourself if you paid attention to that advice.

Ok now I found that advice, not exactly leaping out at you is it now!

As I said in a previous post I don't believe this is a bug as such but rather something to do with my configuration since the behaviour is exactly the same in v7 rc2 as it was in v 6.0.1. Even a configuration issue is unlikely because as far as I am aware nothing has changed, so I guess something changed without my realising it and this is why I am looking for help.

A noob to Linux / SME server I may be but a noob to SQA / limited IT I am not!

I will post a link to this thread in bugs but presumably that will just be registered as a 'bug' and I will still get no answers.

Further more this system has been running with no problems for two years so again I really don't think it is a bug with SME.

What I was after was help in interpreting the information in the logs and perhaps advice on how to obtain more relevant information to help solve the issue.

If I have to be a Linux expert before anyone will take me seriously then.... well what can I say.

Cheers

Andy        

[smeghead]

Check this thread, it might give some clues.  Remember search is your friend:

http://forums.contribs.org/index.php?topic=22814.0

HTH

[abarron]

Thanx smeghead     that's really helpful I will give the advice a go (when I work out how).  I did search but I obviously don't have the hang of this yet 'cos this thread didn't appear.

Thanx again.

Andy.

[abarron]

Ok, again I hope I am not transgressing any ettiqette but I read the thread smeghead pointed me at and have some questions:

Odd behaviour, I disabled virus scanning and set my crontab value to collect mail every minute instead of every 5 minutes and got the following logs.

Messages:

May 14 20:51:44 server esmith::event[32016]: adjust-services=action|Event|email-update|Action|adjust-services|Start|1147636301 727753|End|1147636304 589137|Elapsed|2.861384
May 14 20:54:01 server crond(pam_unix)[32071]: session opened for user root by (uid=0)
May 14 20:54:01 server su(pam_unix)[32073]: session opened for user qmailr by (uid=0)
May 14 20:54:06 server su(pam_unix)[32073]: session closed for user qmailr
May 14 20:54:06 server crond(pam_unix)[32071]: session closed for user root
May 14 20:55:01 server crond(pam_unix)[32089]: session opened for user root by (uid=0)
May 14 20:55:01 server su(pam_unix)[32091]: session opened for user qmailr by (uid=0)
May 14 20:55:04 server su(pam_unix)[32091]: session closed for user qmailr
May 14 20:55:04 server crond(pam_unix)[32089]: session closed for user root


where you can see it starting qmailr every minute, but in

Mailog:

May 14 20:50:04 server fetchmail[32010]: socket error while fetching from mail.uk2.net
May 14 20:50:04 server fetchmail[32010]: Query status=2 (SOCKET)
May 14 20:55:04 server fetchmail[32105]: socket error while fetching from mail.uk2.net
May 14 20:55:04 server fetchmail[32105]: Query status=2 (SOCKET)
May 14 21:00:04 server fetchmail[32241]: socket error while fetching from mail.uk2.net
May 14 21:00:04 server fetchmail[32241]: Query status=2 (SOCKET)
May 14 21:05:04 server fetchmail[32339]: socket error while fetching from mail.uk2.net
May 14 21:05:04 server fetchmail[32339]: Query status=2 (SOCKET)


Why is it I still get a socket error every 5 minutes which is the previous mail collection value but not one every minute?

It is almost as if I have two tasks running one of which invokes a fetchmail (or whatever) that generates a socket error and the other doesn't.

The thread smeghead pointed me at said:

************
oh no, not the Query status=2 (SOCKET) problem. i had to deal with this too. it's not fetchmail, it's clamav that's your problem. (try it if you want by disabling the clamav for a test)

i manually uninstalled clamav (all rpm's of it -think there's three) then reinstalled clamav via the script. then issued the following commands:
/sbin/e-smith/db configuration setprop amavis-ng qmail-queue /var/qmail/bin/qmail-spamc
/sbin/e-smith/signal-event email-update
*************

Could someone explain how I manually uninstall / re-install clamav or point me in the right direction to find out?

Thanks.

Andy

[kruhm]

#im not exactly sure about v7 but v6 would be something like:

rpm -e sme-antivirus
rpm -e clamav-es
rpm -e clamav-es-libs

wget http://sme.swerts-knudsen.com/downloads/SpamFilter/antivirus_install.sh
sh antivirus_install.sh

#you may want to have this thread moved since fetchmail isn't standard

[abarron]

Thanks for trying kruhm but unfortunately this package does not support v7 server.

Thanks anyway.  

[kruhm]

#OK, what about:

rpm -e sme-clamav
rpm -e clamav
rpm -e clamav-db

#Then reinstalling from Yum?

[CharlieBrady]

Thanks for trying kruhm but unfortunately this package does not support v7 server.

clamav has nothing to do with fetchmail mail from your ISP. SME7 already has up to date clamav packages in any case. You appear to be trying to take a step backwards for no good purpose.

[abarron]

Thanks again for your advice kruhm.

Fortunately the problem is resolved so Charlie as I suspected right from the beginning there was no bug to chase.

My mail hosts have upgraded some broken hardware over the course of last night and the problem now 'seems' to be fixed.

Thank you everyone that gave advice I have sure learned a lot, not least that there is an awful lot more for me to learn.      

Cheers
Andy