False rejection of valid sender domain

[#1]

Hi PN,

I have just had a non-delivery error from relay.plus.net saying that the sender domain is invalid:

<*****@maroldt.lu>: host relay.plus.net[212.159.8.107] said: 550
   <*****@topolino.org> sender rejected - invalid domain (no MX or A record)
   (in reply to MAIL FROM command)

The topolino.org domain does however have MX records:

$ dig topolino.org mx

; <<>> DiG 9.10.3-P4-Ubuntu <<>> topolino.org mx
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 56506
;; flags: qr rd ra; QUERY: 1, ANSWER: 2, AUTHORITY: 0, ADDITIONAL: 1

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 512
;; QUESTION SECTION:
;topolino.org.			IN	MX

;; ANSWER SECTION:
topolino.org.		299	IN	MX	0 secure.newtonnet.co.uk.
topolino.org.		299	IN	MX	1 targur.newtonnet.co.uk.

;; Query time: 132 msec
;; SERVER: 8.8.8.8#53(8.8.8.8)
;; WHEN: Fri Nov 17 15:16:16 GMT 2017
;; MSG SIZE  rcvd: 102

I see from the archives that this happened quite a few times before but I haven't found an explanation why.

[#2]

Does the presence of an SPF record which does not include the Plusnet relay server have a relevance here?

topolino.org    text = "v=spf1 include:newtonnet.co.uk a:home.topolino.org a:mailout1.pt.lu a:mailout2.pt.lu a:mailout3.pt.lu a:mailout4.pt.lu -all"

Further why are you relaying via Plusnet's SMTP rather than using the SMTP server associated with the sender's domain?

[#3]

Hi Townman,

Thanks for your thoughts and input on this.

The Plusnet relay is included indirectly via the 'include:newtonnet.co.uk' entry which itself the necessary relay references.

$ dig newtonnet.co.uk txt +short
"v=spf1 mx include:madasafish.com -all" 

$ dig madasafish.com txt +short
"v=spf1 ip4:212.159.14.0/24 ip4:84.93.230.224/27 ip4:62.7.242.1
9 ip4:62.7.242.140 ip4:62.239.224.97 ip4:62.239.224.98 ip4:62.2
9.224.99 ip4:212.159.9.109 ip4:212.159.8.109 ip4:212.159.8.108
p4:212.159.9.108 ip4:213.162.97.76 -all"

The mail for topolino.org domain is hosted on my server (at home) and I relay outgoing mail via Plusnet to avoid any issues associated with sender reputation of customer address space.

What I suspect has happened is that the Plusnet server has been a bit too quick to conclude that the topolino.org domain is invalid, perhaps as a result of a temporary DNS failure. It would be good to hear from Plusnet themselves exactly how this bit of sender validation functions.

[#4]

Cannot put my finger on it, nor prove anything but there were some DNS wobbles during recent days.

When I sought to look at the above domains, I had issues with nslookup (could not find the server), but cannot rule out a local glitch

[#5]

Thanks anyway for looking. It hasn't happened since so I think I'll just have to keep an eye out for it reoccurring. 

[#6]

I've just had this again, with a different sender domain this time:

<*******@gmail.com>: host relay.plus.net[212.159.8.107] said: 550
    <*******@newtonnet.co.uk> sender rejected - invalid domain (no MX or A
    record) (in reply to MAIL FROM command)

As before, the sender domain (newtonnet.co.uk) has all the correct DNS records in place and whilst there is always the possibility of transient DNS failures (lack of resolution response rather than NXDOMAIN) this shouldn't be resulting in false positive rejections like this. Can anyone at Plusnet help out here?

With nothing to fix in the DNS my only resolution is to re-send and hope it goes through (which it always has done so far).

[#7]

I think what you have seen here is part of a small though not insignificant problem that crops up from time to time, although usually connected to the avasins rather than relays. The latest of the latter is topic Plusnet Email servers giving random SMTP 550 errors for inbound emails

I've mentioned this topic in message 21 in that topic https://community.plus.net/t5/Email/Plusnet-Email-servers-giving-random-SMTP-550-errors-for-inbound/... which shows it can also occur when neither the avasins nor the relays are involved.

@bobpullen for reference.

[#8]

Thanks David. As you say it is a small issue, but I am hoping it is small enough to be able to fix.

[#9]

Yes, small but I suspect more prevalent than is/was realised. When Bob searched the inbound email logs there were more instances than the odd post here might suggest (see the other topic), including identification of a start date of 21/10/2017 for the current rejections. Possibly the same with the relays.

If you do encounter more instances please post here to record them. 

[#10]

Can someone from Plusnet please help out with this? It is happening with increasing regularity now and is very frustrating.

[#11]

Please could you post the error message from an up to date instance of this along with the date and time it happened? I hope that information will provide a focus for server log searches.

[#12]

Of course; here are the relevant extracts from my mail log for some recent examples (sender/recipient addresses obfuscated, , and all timestamps are GMT):

Jan  2 10:20:13 rugrat postfix/smtp[16409]: 9F543A29D8: to=<********@electricalsupplies.co.uk>, relay=relay.plus.net[212.159.9.107]:25, delay=14, delays=0.01/0.01/1.1/13, dsn=5.0.0, status=bounced (host relay.plus.net[212.159.9.107] said: 550 <********@newtonnet.co.uk> sender rejected - invalid domain (no MX or A record) (in reply to MAIL FROM command))

Jan  2 10:22:18 rugrat postfix/smtp[16596]: 67743A29D8: to=<********@electricalsupplies.co.uk>, relay=relay.plus.net[212.159.9.107]:25, delay=14, delays=0/0.01/1.1/13, dsn=5.0.0, status=bounced (host relay.plus.net[212.159.9.107] said: 550 <********@newtonnet.co.uk> sender rejected - invalid domain (no MX or A record) (in reply to MAIL FROM command))

Jan  3 08:36:01 rugrat postfix/smtp[25456]: 16ED8A150B: to=<********@pt.lu>, relay=relay.plus.net[212.159.9.107]:25, delay=14, delays=0/0/1.1/13, dsn=5.0.0, status=bounced (host relay.plus.net[212.159.9.107] said: 550 <********@topolino.org> sender rejected - invalid domain (no MX or A record) (in reply to MAIL FROM command))

 In all cases re-sending attempts were eventually successful.

[#13]

Thanks. These examples show rejects also occur from the 212.159.9.107 cluster - the previous ones were from 212.159.8.107.

@bobpullen and @Kelly as with the inbound topic https://community.plus.net/t5/Email/Plusnet-Email-servers-giving-random-SMTP-550-errors-for-inbound/... does a relay log check show these and other instances?

Is DNS giving out information that can be misinterpreted (by internal and external servers)?

[#14]

For completeness here are the DNS records for each domain:

$ dig newtonnet.co.uk mx +short
1 targur.newtonnet.co.uk.
0 secure.newtonnet.co.uk.

$ dig topolino.org mx +short
0 secure.newtonnet.co.uk.
1 targur.newtonnet.co.uk.

$ dig secure.newtonnet.co.uk a +short
81.174.168.158

$ dig targur.newtonnet.co.uk a +short
51.9.7.124

[#15]

Hi, 

Are you still seeing this happening?  We can't see anything suspicious at those times.  There isn't a spike of rejections at the same time which would suggest a wider issue.