HI! It seems that two of my messages to mailman were not forwarded back to me while others came through just fine. One recent example which was successfully delivered to mx2.opensuse.org (see logs below). Message-ID: <16891a5e-fa81-41ec-1d19-f7816ab1e214@stroeder.com> Ciao, Michael. 2021-06-15T18:54:51.636290+02:00 mx1 postfix/cleanup[6452]: 99E3B21493: message-id=<16891a5e-fa81-41ec-1d19-f7816ab1e214@stroeder.com> 2021-06-15T18:54:52.011315+02:00 mx1 postfix/qmgr[1794]: 99E3B21493: from=<michael@stroeder.com>, size=4283, nrcpt=1 (queue active) 2021-06-15T18:54:52.018747+02:00 mx1 postfix/smtpd[6450]: disconnect from mail1.hv.local[10.1.1.7] ehlo=2 starttls=1 mail=1 rcpt=1 data=1 quit=1 commands=7 2021-06-15T18:54:52.141691+02:00 mx1 unbound: [653:0] info: 127.0.0.1 . NS IN 2021-06-15T18:54:52.141769+02:00 mx1 unbound: [653:0] info: 127.0.0.1 . NS IN NOERROR 0.000000 1 525 2021-06-15T18:54:52.163731+02:00 mx1 unbound: [653:0] info: 127.0.0.1 mx1.opensuse.org. A IN 2021-06-15T18:54:52.432899+02:00 mx1 unbound: [653:0] info: 127.0.0.1 mx1.opensuse.org. A IN NOERROR 0.267891 0 50 2021-06-15T18:54:52.788450+02:00 mx1 postfix/smtp[6453]: Untrusted TLS connection established to mx2.opensuse.org[195.135.221.158]:25: TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bi ts) key-exchange ECDHE (P-256) server-signature ECDSA (P-384) server-digest SHA384 2021-06-15T18:54:55.156828+02:00 mx1 postfix/smtp[6453]: 99E3B21493: to=<factory@lists.opensuse.org>, relay=mx2.opensuse.org[195.135.221.158]:25, delay=3.6, delays=0.42/0.06/0.75/2.3, dsn=2 .0.0, status=sent (250 2.0.0 Ok: queued as D7BD48C81)
Could it be related to today's postgres/mirrordb outage? https://progress.opensuse.org/issues/93686 because mailman is one of its users. On 15/06/2021 19.38, Michael Ströder wrote:
It seems that two of my messages to mailman were not forwarded back to me while others came through just fine.
On 6/15/21 10:15 PM, Bernhard M. Wiedemann wrote:
Could it be related to today's postgres/mirrordb outage? https://progress.opensuse.org/issues/93686 because mailman is one of its users.
Yes, probably. In the mean-time the message was delivered. Ciao, Michael.
On 15/06/2021 19.38, Michael Ströder wrote:
It seems that two of my messages to mailman were not forwarded back to me while others came through just fine.
On 15/06/2021 19.38, Michael Ströder wrote:
It seems that two of my messages to mailman were not forwarded back to me while others came through just fine.
On 15/06/2021 23.53, Michael Ströder wrote:
On 6/15/21 10:15 PM, Bernhard M. Wiedemann wrote:
Could it be related to today's postgres/mirrordb outage? https://progress.opensuse.org/issues/93686 because mailman is one of its users.
Yes, probably. In the mean-time the message was delivered.
I noticed three emails of mine that got lost. Per located them in the log entering the systems, but then they went missing in mailman, so I will just resend them. No problem. However, before this I assumed email to be "reliable". Ie, a server passed an email to the next server, but did not delete it till the next hop confirmed reception. This went to the extreme to mount the filesystems holding mail as "sync", not cached, so that if the machine crashed the disk copy would be complete. mailman doesn't seem to be reliable. -- Cheers / Saludos, Carlos E. R. (from 15.2 x86_64 at Telcontar)
On 6/16/21 12:16 PM, Carlos E. R. wrote:
I noticed three emails of mine that got lost. Per located them in the log entering the systems, but then they went missing in mailman, so I will just resend them. No problem. [..] mailman doesn't seem to be reliable.
Without a detailed root-cause analysis it's hard to tell whether mailman was the culprit here. Look again at the ticket Bernhard referenced: https://progress.opensuse.org/issues/93686 Ciao, Michael.
Michael Strc3b6der wrote:
On 6/16/21 12:16 PM, Carlos E. R. wrote:
I noticed three emails of mine that got lost. Per located them in the log entering the systems, but then they went missing in mailman, so I will just resend them. No problem. [..] mailman doesn't seem to be reliable.
Without a detailed root-cause analysis it's hard to tell whether mailman was the culprit here.
Yes, I would say so too - I only looked as far as mailman, I did not look to see what happened in mailman. -- Per Jessen, Zürich (27.2°C) Member, openSUSE Heroes
Carlos E. R. wrote:
However, before this I assumed email to be "reliable". Ie, a server passed an email to the next server, but did not delete it till the next hop confirmed reception.
Why would this incident make you doubt the above? Email works exactly like that, also for us. Postfix on the mailman server confirmed reception, all is good. -- Per Jessen, Zürich (27.2°C) Member, openSUSE Heroes
On 16/06/2021 13.01, Per Jessen wrote:
Carlos E. R. wrote:
However, before this I assumed email to be "reliable". Ie, a server passed an email to the next server, but did not delete it till the next hop confirmed reception.
Why would this incident make you doubt the above? Email works exactly like that, also for us. Postfix on the mailman server confirmed reception, all is good.
That 3 emails got lost. Between postfix and mailman, with no real information, my bets are on mailman being the culprit :-p -- Cheers / Saludos, Carlos E. R. (from 15.2 x86_64 at Telcontar)
Carlos E. R. wrote:
On 16/06/2021 13.01, Per Jessen wrote:
Carlos E. R. wrote:
However, before this I assumed email to be "reliable". Ie, a server passed an email to the next server, but did not delete it till the next hop confirmed reception.
Why would this incident make you doubt the above? Email works exactly like that, also for us. Postfix on the mailman server confirmed reception, all is good.
That 3 emails got lost.
If they got lost, it was not in the email exchange.
Between postfix and mailman, with no real information,
Sufficient information to confirm reception by mailman. -- Per Jessen, Zürich (27.8°C) Member, openSUSE Heroes
On 16/06/2021 13.21, Per Jessen wrote:
Carlos E. R. wrote:
On 16/06/2021 13.01, Per Jessen wrote:
Carlos E. R. wrote:
However, before this I assumed email to be "reliable". Ie, a server passed an email to the next server, but did not delete it till the next hop confirmed reception.
Why would this incident make you doubt the above? Email works exactly like that, also for us. Postfix on the mailman server confirmed reception, all is good.
That 3 emails got lost.
If they got lost, it was not in the email exchange.
Between postfix and mailman, with no real information,
Sufficient information to confirm reception by mailman.
I mean "information by me" for making an informed impression :-) You have information and knowledge. You are confirming what I say, you also think mailman is the one. -- Cheers / Saludos, Carlos E. R. (from 15.2 x86_64 at Telcontar)
On 6/16/21 2:07 PM, Carlos E. R. wrote:
You are confirming what I say, you also think mailman is the one.
Life is a bit more complex. Even if a message got lost after being accepted by mailman it does not necessarily mean that the mailman devs could do anything about it. Basically during conditions like full file systems, half-working databases or similar, there is no system integrity at all. Ciao, Michael.
On 16/06/2021 14.38, Michael Ströder wrote:
On 6/16/21 2:07 PM, Carlos E. R. wrote:
You are confirming what I say, you also think mailman is the one.
Life is a bit more complex.
Even if a message got lost after being accepted by mailman it does not necessarily mean that the mailman devs could do anything about it.
Basically during conditions like full file systems, half-working databases or similar, there is no system integrity at all.
Well, the theory says that mailman should not accept mail under those conditions. Ie, -receive the email, keep the connection open with the sender system -store it on disk, update databases, all that. -failed? tell the sender system that mail was not accepted, internal trouble. -success? tell the sender mail was accepted and close connection. :-) -- Cheers / Saludos, Carlos E. R. (from 15.2 x86_64 at Telcontar)
Carlos E. R. wrote:
On 16/06/2021 14.38, Michael Ströder wrote:
On 6/16/21 2:07 PM, Carlos E. R. wrote:
You are confirming what I say, you also think mailman is the one.
Life is a bit more complex.
Even if a message got lost after being accepted by mailman it does not necessarily mean that the mailman devs could do anything about it.
Basically during conditions like full file systems, half-working databases or similar, there is no system integrity at all.
Well, the theory says that mailman should not accept mail under those conditions.
In principle perhaps, but in practice mailman is not an MTA and does not behave like one. -- Per Jessen, Zürich (26.1°C) Member, openSUSE Heroes
On 16/06/2021 21.10, Per Jessen wrote:
Carlos E. R. wrote:
On 16/06/2021 14.38, Michael Ströder wrote:
On 6/16/21 2:07 PM, Carlos E. R. wrote:
You are confirming what I say, you also think mailman is the one.
Life is a bit more complex.
Even if a message got lost after being accepted by mailman it does not necessarily mean that the mailman devs could do anything about it.
Basically during conditions like full file systems, half-working databases or similar, there is no system integrity at all.
Well, the theory says that mailman should not accept mail under those conditions.
In principle perhaps, but in practice mailman is not an MTA and does not behave like one.
My thought exactly :-) -- Cheers / Saludos, Carlos E. R. (from 15.2 x86_64 at Telcontar)
On 16/06/2021 14.38, Michael Ströder wrote:
On 6/16/21 2:07 PM, Carlos E. R. wrote:
You are confirming what I say, you also think mailman is the one.
Life is a bit more complex.
Even if a message got lost after being accepted by mailman it does not necessarily mean that the mailman devs could do anything about it.
Basically during conditions like full file systems, half-working databases or similar, there is no system integrity at all.
Well, the theory says that mailman should not accept mail under those conditions. Ie, -receive the email, keep the connection open with the sender system -store it on disk, update databases, all that. -failed? tell the sender system that mail was not accepted, internal trouble. -success? tell the sender mail was accepted and close connection. :-) -- Cheers / Saludos, Carlos E. R. (from 15.2 x86_64 at Telcontar)
participants (5)
-
Bernhard M. Wiedemann -
Carlos E. R. -
Carlos E. R. -
Michael Ströder -
Per Jessen