timeout on 127.0.0.1:10022 while reading input attribute value

[kiddx]

Prox is working fine for us except from one client who sends us mail from a large hospital group, we used to get their email no problem. everything else seems to be working okay, mail is flowing fine.

[TR]
[TD]timeout on 127.0.0.1:10022 while reading input attribute value[/TD]
[/TR]
[TR]
[TD]Nov 2 08:53:42[/TD]
[TD]smtpd[/TD]
[TD]warning: problem talking to server 127.0.0.1:10022: Connection timed out[/TD]
[/TR]
[TR]
[TD]Nov 2 08:55:23[/TD]
[TD]smtpd[/TD]
[TD]warning: timeout on 127.0.0.1:10022 while reading input attribute value[/TD]
[/TR]
[TR]
[TD]Nov 2 08:55:23[/TD]
[TD]smtpd[/TD]
[TD]warning: problem talking to server 127.0.0.1:10022: Connection timed out



then proxmox sends back a 451 4.3.5 Server configuration problem. I did some reading and came up with a postgrey database error. i rebooted prox, stop/started the services, but since this isnt system wide I dont know that the issue is on my end. Since its a large hospital I would figure the problem is not on their end, but not sure what else to do. I even tried turning off greylisting .... I also whitelisted the IP address, which we have a handful of whitelisted IPs in the rules , and still doesnt come through.[/TD]
[/TR]

 

[kiddx]

Okay, have a few more with the same issue that we didnt know about. In reading this is a postgrey corruption issue, so i was hoping to ssh in and move the postgrey db out, but doesnt look like its where the default is, searching the whole box for *grey* doesn't seem to show me anything ... anyone know where i can find the postgrey db and reset it?

thx

 

[tom]

looks like your filter service is stopped. check all your services.

 

[kiddx]

from the gui all services are running, mail is flowing fine but we now have 3 clients who have sent us mail and that did not fix it. We thought it was just one initially so didnt think too much about it, but now today its 2 more ... i have rebooted prox stopped/started services in the GUI ... i noticed when i restarted policy the SMTP daemon went down and didnt come back up, but hitting start worked just fine and we were good to go... the few people that have this error seem to say the postgrey db is corrupt and to move it out and restart the box (or postgrey) so it builds a new DB ...

 

[dietmar]

Proxmox does not use 'postgrey' (greylisting is done by the 'proxpolicy' daemon)

 

[kiddx]

Okay, after a little digging we found an item in the postfix conf file that says to use the postgrey, we commented that line out and hopefully tomorrow we will verify its working okay.

Thanks

 

[dietmar]

Okay, after a little digging we found an item in the postfix conf file that says to use the postgrey

What???

 

[kiddx]

heh, I did some digging around about postgrey, and found mention of editing main.cf for postfix with check_policy_service inet:127.0.0.1:10022

This was my error we started getting a few days ago. I forget where the conf file was, it was in /etc , we grepped the whole system for the check_policy_service and found it.

After removing that, the error went away and one of the emails that was giving me this error was delivered, so I think we are going to be in good shape from here. Just odd that it just suddenly started querying the postgrey system, even though nothing has changed, to my knowledge in many months.

 

[kiddx]

Okay, we have isolated the issue but not sure what to do about it... this morning the same problem came back, when we went into /etc/postfix/main.cf there is check_policy_service = inet:127.0.0.1:10022 <--- this is the exact area we get in tracking center.

I commented this out again, then when I went into the web gui. I turned off greylisting and turned it back on and came back to this file and it was un-commented again. So it appears to me that turning on greylisting re-enables the check_policy_service within postfix for some reason. This is whats causing our greylisting problems. Im going to switch to mx.fakemx.net soon but would like to keep this on if possible...

anyone have any clues? At this point I would like to just clear out the greylist database and let it rebuild. This seems to fix this error with postgrey, so Im trying to find some commonality between what proxmox is doing and what postfix is doing with postgrey...

Thanks

 

[dietmar]

This seems to fix this error with postgrey, so Im trying to find some commonality between what proxmox is doing and what postfix is doing with postgrey...

Again, Proxmox Mail Gateway does not use postgrey!

 

[kiddx]

I understand that proxmox doesnt, however in my logs I get an error connecting to 127.0.0.1:10022. this is listed under the postfix/main.cf check_policy option. If I comment it out then everything works okay. if I turn greylisting back on in the proxmox GUI, this config file gets changed and turned back on and doesnt work again.

When I google this error it continuously refers to postgrey and the postgrey database. So, while it was not postgrey, basically our greylisting database was corrupted.

We gave up and just wiped proxmox and rebuilt a new one and imported the config ... so far so good.

We never could figure out if this was occuring for all greylisting, because we get emails from all kinds of people and only noticed 2 particular domains, which were working a couple of weeks ago and should have already been in the greylist database ... in any event, hopefully tomorrow will be better.

thanks

 

[dietmar]

I understand that proxmox doesnt, however in my logs I get an error connecting to 127.0.0.1:10022. this is listed under the postfix/main.cf check_policy option.

Yes, that is the connection to our policy daemon (proxpolicy) - it is a bad idea to remove that!

If I comment it out then everything works okay. if I turn greylisting back on in the proxmox GUI, this config file gets changed and turned back on and doesnt work again.

Simply disable greylisting on the web interface If you do not want greylisting.

When I google this error it continuously refers to postgrey and the postgrey database. So, while it was not postgrey, basically our greylisting database was corrupted.

I doubt that this is the reason.

 

[kiddx]

Thanks for the help everyone, unfortunately the boss came and said to remove it. I tried to explain it must be an issue with the senders email (they are sending https redirects) but since our attorney now went and got a yahoo address and he gets them all he came back and said it must be something in our end... I never was able to find the problem with this 127.0.0.1:10022 for just a few particular domains that are coming in, it just happens one of them is from a big new client.

Taking off greylisting works, and uncommenting that line, which has no effect, does have some effect because as soon as I do that we are able to get all the mail and this error goes away. this is a brand new 2.6 install and we thought with a new installation (we did export/import the config from a couple weeks ago) everything would be okay, but nope, today i had to comment out the check_policy_service because it keeps getting uncommented.

I have no idea what is supposed to be listening on 127.0.0.1:10022 other than check_policy_service inet/127.0.0.1:10022 in the postfix main.cf which isnt used, but my proxmox thinks so I guess.

Appreciated the help though!

Thx

 

[donluis]

I encountered this Error last week the first time (one sender in hundred thousands) and found out, that he had a broken SPF-entry in his DNS.
So i believe that the broken DNS SPF-Line produces a timeout in proxmox.