• If you are still using CentOS 7.9, it's time to convert to Alma 8 with the free centos2alma tool by Plesk or Plesk Migrator. Please let us know your experiences or concerns in this thread:
    CentOS2Alma discussion

localhost smtp broken after plesk 9 upgrade (with solution)

E

eMediaSales

Guest
Hope this helps someone else, I spent quite a few hours here then finally came across the clue I needed to get this fixed. Essentially mail from localhost should always be permitted (eg: localhost SMTP to port 25). However, I was consistently getting the following when emailing from localhost:

553 sorry, that domain isn't in my list of allowed rcpthosts (#5.7.1)

FIX: Replace /var/qmail/bin/tcp-env with a known good copy from a Plesk 8 distribution.

The issue is that the tcp-env shipped with Plesk9 is completely broken. Replacing /var/qmail/bin/tcp-env with a known working copy from a Plesk 8.6 server resolves the issue. I tested locally (telnet localhost 25) and am now able to send outbound email. Also tested from a remote server to ensure that it doesn't create an open relay. - eg: relay should always be allowed locally (localhost whitelist) where as non-rcpthost emails arriving on port 25 should require SMTPAUTH.

FWIW -
James
 
This is extremely useful to know. Thank you,

It would be nice for this topic to be made sticky until this bug is fixed.

Faris.
 
yep, would you be nice enough to attach a working copy of this file? :)
 
The easiest way to get a tpc-env for your distro is by going to the following URL:

http://autoinstall.plesk.com/PSA_8....psa-qmail-1.03-cos5.build86080930.08.i586.rpm

make sure to select the right OS (the example above includes CentOS 5, i386 / 686) and download the latest build you can find.

Use something like midnight commander to open the RPM and copy tcp-env to the right location and you'll have working mail again.

Now, if only someone could tell me how to fix Horde, I'd be a happy camper
 
hallelujah!. Thanks a million, all fixed for me too!

And, btw, 7-zip allows you to read .rpm's on windows. Also make sure you're downloading the qmail rpm, not plesk's base rpm -- tcp-env isn't in there. :)

wshew ... what a pain!
 
Ditto

I too have been receiving complaints about the 553 error since we upgraded. I've tried the change recommended regarding restoring tcp-env. I'll follow-up if I get reports that it didn't work.
 
Not so good

I started receiving a large number of reports that email got broken altogether so I reversed this update. Back to the drawing board...
 
Same here, after pulling the tcp-env from a version 8.6, qmail blew up. Restored all files and back to square 1. Side note that most of the problems I have seen w/ 9.0 is due to an upgrade from previous versions. All fresh installs have been smooth (so far...). Hopefully there will be an updated tcp-env from Parallels soon.
 
Hi!
I'm getting this bounce from some servers since upgrading to 9.0.0 (upgrade to 9.0.0.2 hasn't worked so far).

Remote host said: 504 <server>: Helo command rejected: need fully-qualified hostname

Do you think this might be the same cause?
Thanks heaps!
G
 
Trying now to get under the hood on qmail

Perhaps somebody knows the answer to this. I spent some of my good money on a support ticket from Parallels. They logged into my server (a CentOS 4.5 box) and looked around. My clients have been reporting gobs of 553 errors. But they claimed to have found only one instance of the error in the log files and are now throwing it back at me to reproduce the error. Nice!

So since it's on me, I thought that I'd see if I could at least look for the evidence myself. I've found /usr/local/psa/var/log/maillog so far. In looking at that, Parallels is right. I see a lot of stuff in there but very very few entries recording the 553 rcpthosts error. I know they're happening, so where can I find the evidence? Do I need to turn some extra logging on? Or is there another place to search? I need to get to the bottom of this before my clients run out of patience....
 
/usr/local/psa/var/log/maillog is the only current mail log. but you may have maillog.1.gz or similar, and maillog.processed (possibly) with older entries.

If it isn't in any of those, it may be a case of the user exagerating - which does happen.

Or it maybe something else entirely -- maybe someone else will suggest something.

Faris.
 
Hi!
I'm getting this bounce from some servers since upgrading to 9.0.0 (upgrade to 9.0.0.2 hasn't worked so far).

Remote host said: 504 <server>: Helo command rejected: need fully-qualified hostname

Do you think this might be the same cause?
Thanks heaps!
G

Hi,

Make sure that current hostname of server resolves to its ip address and is a FQDN. You can find what current hostname of server it using the following commands:
# hostname -- hostname of server
# cat /var/qmail/control/me -- hostname send in qmail greeting

You can change server hostname via Plesk panel in server preferences.
 
Well I've finally had enough

It's been three weeks since I "upgraded" from 8.6 to 9.0 . Since that update customers have been having constant problems with their outgoing email. I've been able to narrow it down to an issue with emails with large attachments (< 2MB) seem to fail more often than small ones. I even purchased a support ticket for Parallels to work on the issue.

To date, I've been underwhelmed with the response. Half the time, it seems like the support staff ignores the message threads in the ticket and asks me questions that were already answered or just stalls me. I've read the forums in the past and knew that updates were always a roll of the dice. Now I've been personally burned by the lower quality of the releases.

So tonight I'm going to be extracting and moving some key accounts to another server that I have that's still running 8.6 . It's not really what I wanted to do but in order to stay in business I have to. Add me to the list of people that are ready to head to the exits.
 
Here's one that one of you fine folks might know, or have seen.
when upgrading to 9.0 i got the "email of death" from the auto-installer :(
This was back in Dec 2008. I contacted Plesk support and sure enough, they got the server going again for me and did so at no charge.. something that actually impressed me quite a bit.
now comes the weird part.
I pop emails off of my server via spamcop and i started seeing a few emails that had no sender, no subject line, and no other discernible information, but if i clicked on the email in the IMAP section, lo and behold, there's what i can only describe as a spammer's failed attempt at sending an HTML email(spam) out...

OK, so there's just a few, and I've seen that type behavior from broken spam messages before, so it's all good, i delete them because spamcop won't parse them automatically and i go on about my business...

well 2 days later, i see a couple of hundred that are all the same exact way, so again, i laugh at the idiot spammer and go on about things as normal.

well my laughter was seriously misplaced... or at least where i had it directed.. i sent myself an email in a CC so i could keep a record of the message and there i was, no header, no from, no to, and the broken html **** inside an email that i knew for a fact came from a system where nothing was changed...

now, i'm no longer laughing.....
one of my domains on that system has the kayako support suite installed on it, and uses piping, but most emails sent to that shows a failure to create a ticket because no recipient is specified... great, so much for supporting my customers.
were it not for one client who happened to have my aim, i wouldn't have thought to even check that.. i thought it was just a nice easy day in support....

I saw a couple of threads that talked about a command line option to fix qmail after a broken Plesk 9 upgrade, but for the life of me i can't find it now, and one of our busier systems just got it's "email of death" sent out too.

any help here would be appreciated as this issue has by far exceeded the boundaries of being funny by any stretch of the imagination.

Thanks in advance..
Bill
 
same problem here

Hi,
i have the same problem on a debian server. A reboot fixes the problem temporarily until the next morning.
Something seems to break qmail at night...

Markus
 
Hi,
i have the same problem on a debian server. A reboot fixes the problem temporarily until the next morning.
Something seems to break qmail at night...

Markus

Hi Markus,
Thanks, at least i know it's "not just me"...
unfortunately, i don't have the luxury of rebooting the server every day, if i end up having to do that, then i might as well just -halt it and be done with it...

hopefully someone will see our messages and will be able to point us in the right direction. :)

well, i found the edit button rofl..
 
Last edited by a moderator:
*resolved*

Here's one that one of you fine folks might know, or have seen.
when upgrading to 9.0 i got the "email of death" from the auto-installer :(
This was back in Dec 2008. I contacted Plesk support and sure enough, they got the server going again for me and did so at no charge.. something that actually impressed me quite a bit.
now comes the weird part.
I pop emails off of my server via spamcop and i started seeing a few emails that had no sender, no subject line, and no other discernible information, but if i clicked on the email in the IMAP section, lo and behold, there's what i can only describe as a spammer's failed attempt at sending an HTML email(spam) out...

[snipped]

I saw a couple of threads that talked about a command line option to fix qmail after a broken Plesk 9 upgrade, but for the life of me i can't find it now, and one of our busier systems just got it's "email of death" sent out too.

any help here would be appreciated as this issue has by far exceeded the boundaries of being funny by any stretch of the imagination.

Thanks in advance..
Bill

bad to be "quoting" myself, but i didn't see where to "edit" the post...
(well, not till after i wrote this... the edit button is grey'd out, but evidently works still)

here's the link to a KB article that helped me fix my email issue:

http://kb.odin.com/en/5847

Thanks
Bill
 
Last edited by a moderator:
Yesterday afternoon I applied the latest 9.0.1 update. Didn't seem to fix the 553 problem and could have possibly made things worse as I'm getting reports today from customers of slow outbound mail and other strange errors. God how I wish I hadn't left 8.6...
 
Back
Top