Plesk 10.2 - Trouble sending E-Mails to "freenet.de"

[TMASBS]

My kernel is 2.6.32-042stab055.12.

depmod -a returns the same errors:
WARNING: Couldn't open directory /lib/modules/2.6.32-042stab055.12: No such file or directory
FATAL: Could not open /lib/modules/2.6.32-042stab055.12/modules.dep.temp for writing: No such file or directory

Should I upgrade my kernel?

Thanks!

 

[HostaHost]

Upgrade and ensure the new one comes with all the associated libraries it should have, or get the library package for the kernel you're currently running and extract/install it. I'm only familiar with rhel/centos so I'm not sure what type of kernel or platform you're running as the name looks unfamiliar.

 

[Pierre-LaurentS]

Hi all,

I'm currently stuck with the very same issue on a soon to be migrated platform (but not ready yet) :

# cat /usr/local/psa/version
10.2.0 CentOS 4.2 1011110331.11

# uname -ar
Linux 2.6.18-028stab099.3 #1 SMP Wed Mar 7 15:20:22 MSK 2012 i686 i686 i386 GNU/Linux

# rpm -qa |grep qmail
psa-qmail-1.03-cos4.build1011110330.18
psa-qmail-rblsmtpd-0.88-cos4.build1011110330.18

Do you think it's possible to have the workaround applied on this platform ?

Thank you very much in advance.

 

[HostaHost]

The workaround requires being able to enable the ipv6 kernel module and assigning a bogus ipv6 address to the primary interface; if your kernel does not include the relevant modules, it's impossible to work around since Parallels doesn't appear to have any interest in fixing the issue.

 

[TMASBS]

I've tried to upgrade and run updates but that did not fix the problem.

Could someone from Parallels take a look at this issue?

Thanks!

 

[HostaHost]

Just a thought; if you have an smtp relay host elsewhere that you can open up relaying on for traffic from your server with the problem, you could probably create a gmail.com:#.#.#.# entry in the /var/qmail/control/smtproutes file with the IP of your other server and just force all the email for gmail to go through the other server as a workaround while we wait another year for Parallels to fix this.

 

[TMASBS]

I switched from qmail to postfix and the problem is solved.
Thank you for the quick replies!

 

[JTRipper]

The problem in that qmail is compiled with IPv6 support and tries to get DNS information with help of IPv6 primitives. For fixing this issue you can use qmail-remote.moved from Plesk 10.1 version without IPv6 support implemented. Also as possible workaround you can enable IPv6 on network interface without assigning real IPv6 address. It should help.

Would you be so kind and provide us a link to this file ? There is no file here anymore from 10.1 version ftp://download1.swsoft.com/Plesk/PP10/10.1.0/

 

[KristianR]

We published a workaround in our knowledgebase here:
http://www.web-vision.de/support/in..._unavailable-430---plesk-qmail-fehler---error

Unfortunately it's only available in german now, but you might use google translate.

Regards,

Boris

This worked for me. Thanks Boris

 

[IgorG]

Would you be so kind and provide us a link to this file ? There is no file here anymore from 10.1 version ftp://download1.swsoft.com/Plesk/PP10/10.1.0/

Look at http://64.131.90.31/PSA_10.1.1/dist-rpm-xxxxxxx/opt/maildrivers/ This file is a part of one of qmail packages.

 

[Trevor L]

Temp fixed using smtproutes file did work.

Hi guys,

Server,Plesk version I have.
Plesk 10.2.0
CentOS release 5.6 (Final)
kernel-2.6.18-238.el5

I'm wondering if upgrading to Plesk 10.4.4 will solve this problem of "System resources temporarily unavailable. (#4.3.0)" bounce backs? My Plesk says when logging in "Upgrade to Parallels Plesk Panel 10.4.4 is available." Will upgrading fix the problem?

We currently had problems sending email to @gmail.com and to fix this I did use what people said on this thread for a temp fix that worked. Read below if you need to do this too for now. It's a simple temp fix.

Logged into server SSH using putty.exe
ran command:
nano /var/qmail/control/smtproutes

I added to the smtproutes file on it's own line.
gmail.com:173.194.77.26

I got the IP by going to mxtoolbox.com and typing in gmail.com and getting the first MX IP.

I restarted SMTP Server (QMail) within Plesk by going to Tools & Utilities > Services management > Restart SMTP Server (QMail)

Since i had emails in the queue still: Plesk > Settings > Mail Server Settings > Tab Mail-Queue
I ran this command to force send queued emails that didn't go already.
kill -ALRM `ps ax | grep qmail-send | grep -v grep | awk '{print $1}'`

After running the command many gmail.com emails left the queue. I had one going to my gmail address that was in the queue and I did receive it.

 

[HostaHost]

I'm wondering if upgrading to Plesk 10.4.4 will solve this problem of "System resources temporarily unavailable. (#4.3.0)" bounce backs? My Plesk says when logging in "Upgrade to Parallels Plesk Panel 10.4.4 is available." Will upgrading fix the problem?


I added to the smtproutes file on it's own line.
gmail.com:173.194.77.26

10.4.4 will not fix the problem. Your fix was a good idea though for a temporary workaround; only issue there is if google changes the IP you won't pick it up.

 

[Trevor L]

Yes, you are right. I was hoping upgrading Plesk would fix the issue all together. I'll have to check gmail.com MX records each day to see how often they change.

If I could find a fix that someone done using, yum command I may try it if they had CentOS.

Maybe I just need to upgrade to a new server that has the newest version of CentOS with newest verison of Plesk and it would be fix. All in all if people using the newest Plesk version can't send email to gmail.com users it's a huge problem now.

 

[Trevor L]

Complete fix after another try at enabling IPV6

Hi guys,

Just wanted to give an update. Before I posted to this Thread I gave what BunyaminV said to do by enabling IPV6 yet it didn't work after doing what he said. Gave me CRITICAL : [ipv6_test] Kernel is not compiled with IPv6 support when running the command: service network restart

I went back and thought I would see what did I miss before. Looking into it i noticed I had to do something else that BunyaminV left out. I'm posting what worked for me below if you tried enabling IPv6 and couldn't get it to work. This fixed emails going to gmail.com and other emails that bounced back with
System resources temporarily unavailable. (#4.3.0)
I'm not going to try again; this message has been in the queue too long.

So I log into the server with SSH using putty.exe
I typed in the command and made changes to NETWORKING_IPV6=no and made it =yes:
# nano /etc/sysconfig/network
NETWORKING_IPV6=yes

I typed in this command and added to the end of the file these two lines.
# nano /etc/sysconfig/network-scripts/ifcfg-eth0
IPV6INIT=yes # Enables IPv6 addressing on interface.
IPV6_AUTOCONF=yes # Enables EUI 64 auto-addressing.

This is where I went wrong before. My file looks like this:
# nano /etc/modprobe.conf
alias eth0 e1000e
alias eth1 e1000e
alias scsi_hostadapter 3w-9xxx
alias scsi_hostadapter1 ahci
alias scsi_hostadapter2 usb-storage
alias net-pf-10 off
alias ipv6 off
options ipv6 disable=1

As you can see above the original file has the last 3 lines as off, off, =1. I changed these 3 lines with this
#alias net-pf-10 off
#alias ipv6 off
options ipv6 disable=0

I comment out the first 2 lines and changed options ipv6 disable=1 to =0. BunyaminV said to put alias ipv6 on but didn't say anything else to do in the file. I can clearly see now I should of seen the disable=1 but I didn't the first time. Also a reboot was needed for the stuff to finally work.

I ran the command
# service network restart

And still got the error "CRITICAL : [ipv6_test] Kernel is not compiled with IPv6 support"

I rebooted the server and crossed my figures it would come back online. Domain was down and putty.exe went dead. A Minute or two later of keep trying the websites and logging back into SSH the server come back online.
# shutdown –r now

I ran the command again and this time the error was gone.
# service network restart

On a side note: I removed the smtproutes I setup before I started and sent an email to my gmail and it just sat in the queue for 13min. After I setup the IPv6 everything started to work.
# nano /var/qmail/control/smtproutes
gmail.com:173.194.77.26 # I removed this line

Now of course email that I send from now on doesn't go out of the mail queue right away it takes up to 5 seconds to send to gmail but it now sends out. I can confirm as well that other domains I had the problem with now get email our domain sends too.
I also did run
# kill -ALRM `ps ax | grep qmail-send | grep -v grep | awk '{print $1}'`
That after I did sent out what was in the queue. It just takes a bit of time before it really does work so give it sometime to do it's work. Like at least a Minute.

Thanks to BunyaminV and others for putting me on the right track I just missed a few settings to get it to work completely.

My server info is:
# cat /etc/redhat-release
Server: CentOS release 5.6 (Final)

#uname -a
Linux 2.6.18-238.el5 #1 SMP Thu Jan 13 15:51:15 EST 2011 x86_64 x86_64 x86_64 GNU/Linux

Plesk 10.2.0

 

[KristianR]

Rebooting the hardware node to enabled IPv6 isn't always an option for obvious reasons. The quickest solution to this problem was posted by Igor and Boris earlier in the thread:
http://forum.parallels.com/showpost.php?p=454947&postcount=16
http://forum.parallels.com/showpost.php?p=616590&postcount=27

Since Boris site is in German, here is the outline:
wget http://autoinstall.plesk.com/PSA_10...er-10.10.1-cos5.build1010110120.18.x86_64.rpm
wget http://autoinstall.plesk.com/PSA_10...qmail-1.03-cos5.build1010110112.13.x86_64.rpm

rpm -Uvh plesk-mail-qc-driver-10.10.1-cos5.build1010110120.18.x86_64.rpm --force --nodeps
rpm -Uvh plesk-mail-qc-driver-10.10.1-cos5.build1010110120.18.x86_64.rpm --force
rpm -Uvh psa-qmail-1.03-cos5.build1010110112.13.x86_64.rpm --force
/etc/init.d/qmail restart
/etc/init.d/xinetd restart

Check the queue:
/var/qmail/bin/qmail-qstat

Force send if necessary:
kill -ALRM `pgrep qmail-send`

Check the queue again:
/var/qmail/bin/qmail-qstat

Once the queue is empty, leave the temporary fix in place until parallels patches, or switch the postfix as a perm solution: http://kb.parallels.com/en/5801

A much cleaner and less time consuming method of fixing this problem without imposing an outage to customers...

 

[wkrick]

From what I've read online, postfix is more actively developed than qmail, so it seems that it should probably be the default choice anyway.

Are there any drawbacks or known issues with postfix that would make switching not a desirable option?

Are there any configuration issues that need to be considered with other modules? For example, SpamAssassin.

 

[HostaHost]

From what I've read online, postfix is more actively developed than qmail, so it seems that it should probably be the default choice anyway.

Are there any drawbacks or known issues with postfix that would make switching not a desirable option?

Are there any configuration issues that need to be considered with other modules? For example, SpamAssassin.

One MTA versus another is like a religious debate; no winners. qmail doesn't have heavy development mostly because it does exactly what it was designed to do and that's it. It's an extremely stable and extremely secure mta when used as intended. This whole issue has come up because Parallels patched qmail with some third party patch to make it ipv6-aware and hasn't bothered to put a second of work into making sure it works correctly in all scenarios; and also hasn't bothered to put a second of work into fixing it once they were made aware of the problems a year ago.

It probably is a good idea to move towards postfix though, for those who plan to keep using Plesk, as it seems that's what Parallels prefers and what gets the new features (for example, the recently added sending of outbound email on the customer's IP instead of the server's primary IP only works on postfix).

We've been sticking with qmail simply because it's what we know well and it's secure. We recently started running dual stack on our network so the ipv6 issue became a non-issue, otherwise we'd probably have to start learning postfix sooner.

 

[wkrick]

It probably is a good idea to move towards postfix though, for those who plan to keep using Plesk, as it seems that's what Parallels prefers and what gets the new features (for example, the recently added sending of outbound email on the customer's IP instead of the server's primary IP only works on postfix).

I'd definitely like to set that up. Is this something that you have to manually configure for each domain? Could you point me at some documentation for this feature (if you know of any)?

 

[KristianR]

I'd definitely like to set that up. Is this something that you have to manually configure for each domain? Could you point me at some documentation for this feature (if you know of any)?

To switch over to postfix, refer to the kb article: http://kb.parallels.com/en/5801

 

[HostaHost]

I'd definitely like to set that up. Is this something that you have to manually configure for each domain? Could you point me at some documentation for this feature (if you know of any)?

LOL anyone reading this thread will get a kick out of this. I went to find the docs on how to do that for you, and did:

http://download1.parallels.com/Ples...nistration-guide/index.htm?fileName=70212.htm

however, doing this has a prerequisite of the server having IPv6 configured....