No delivered mail appears in tracking center

[Dark26]

Bonjour,

In the tracking center in my personnal mail gateway no problem with the tracking center :




But in our production environnement ( cluster ), we only have this :



no accepted delivered mail. And for some domain we filter, there is not entry at all. But the spam quarantine is working

Maybe we have done (not sure) some changement for the logging /rotate system,

Do you have some ideas where to look for the resolving the problem ?

Thanks

 

[Dark26]

After investigation, i think we found the problem :

on the postfix configuration we enable the parameter : enable_long_queue_ids

so instead of having :

2025-09-25T00:01:18.236323+02:00 mailgateway postfix/smtp[84617]: 81CE38C8: to=<xxxx@yyyy.fr>, relay=172.16.4.198[172.16.4.198]:25, delay=0.72, delays=0.08/0.2/0.29/0.16, dsn=2.0.0, status=sent (250 2.0.0 Ok: queued as 4cX9ly1T32zKBcY)

we have something like that :

2025-09-25T13:36:12.466175+02:00 gg02 pmg-smtp-filter[2005052]: 4000368D5292BB51FB: accept mail to <rrr.rrrr@rrr.fr> (4) (rule: default-accept)
2025-09-25T13:36:12.470492+02:00 gg02 postfix/lmtp[2000844]: 4cXWrC4l0Gz5wnK: to=<rrr.rrrr@rrr.fr>, relay=127.0.0.1[127.0.0.1]:10024, delay=0.8, delays=0.03/0/0.04/0.73, dsn=2.5.0, status=sent (250 2.5.0 OK (4000368D5292BB51FB))
2025-09-25T13:36:12.496045+02:00 gg02 postfix/smtp[2004819]: 4cXWrD2zLhz5wnq: to=<rrr.rrrr@rrr.fr>, relay=mmm.mm.mm.fr[10.10.10.10]:25, delay=0.08, delays=0.05/0/0/0.03, dsn=2.0.0, status=sent (250 2.0.0 Ok: queued as 4cXWrD3Sq8z4x5fq)

according to https://git.proxmox.com/?p=pmg-log-...9b57008f738b0165ec9ff189214c8dfc4932d00#l2157

maybe something is wrong here ( or else were):

 fn parse_qid(data: &[u8], max: usize) -> Option<(&[u8], &[u8])> {
2194     // to simplify limit max to data.len()
2195     let max = max.min(data.len());
2196     // take at most max, find the first non-hex-digit
2197     match data.iter().take(max).position(|b| !b.is_ascii_hexdigit()) {
2198         // if there were less than 5 return nothing
2199         // the QID always has at least 5 characters for the microseconds (see
2200         // http://www.postfix.org/postconf.5.html#enable_long_queue_ids)
2201         Some(n) if n < 5 => None,
2202         // otherwise split at the first non-hex-digit
2203         Some(n) => Some(data.split_at(n)),
2204         // or return 'max' length QID if no non-hex-digit is found
2205         None => Some(data.split_at(max)),
2206     }
2207 }

because like the comments, we activate the enable_long_queue_ids, but the test doesn't seems to work. maybe because the value isn't with hexacharater ? (b.is_ascii_hexdigit)

 

[Stoiko Ivanov]

on the postfix configuration we enable the parameter : enable_long_queue_ids

This is not supported by the log-tracker (and there are not current plans to change that):

https://bugzilla.proxmox.com/show_bug.cgi?id=2091

I hope this helps!

 

[Dark26]

This is not supported by the log-tracker (and there are not current plans to change that):

https://bugzilla.proxmox.com/show_bug.cgi?id=2091

I hope this helps!

Now we know why it doesn't work, but it doesn't help for the tracking of the message.

 

[Stoiko Ivanov]

Why do you need to set enable_long_queue_ids? - without this setting the tracking center would work as intended.

 

[Dark26]

Why do you need to set enable_long_queue_ids? - without this setting the tracking center would work as intended.

Without this parameter , the id of the email is too short when you have a lots of mail, and a lots of logs, you can have two differents mails with the same ID.

PMg is the "man in the middle". We have another appliance for filtering before PMG with logs , and another filtering before the final local MTA with logs too.

In order to "track" the mail , when for example it's block somewhere for some reason, we need to have the same level of configuration the logs, or its's impossible to follow the trace between all the logs, and it's a nighmare.

 

[Stoiko Ivanov]

Without this parameter , the id of the email is too short when you have a lots of mail, and a lots of logs, you can have two differents mails with the same ID.

did this ever happen for you in practice?

and - the IDs are local to the system - yes you see them in the sending server's logs, and when sending mails further in those servers logs as well... - but to get the complete picture you already need to look at all system's mail-ids... (even im PMG you have 3 ID's for each processed mail (postfix when it comes in, pmg-smtp-filter, postfix when it goes out...)