Hello,

czw., 1 paź 2020 o 16:45 <djosip+n...@linuxpages.net> napisał(a):

> > Yes, it is true to some extent.
>
> Hm, I believe it happens in all cases.
> In what case it wouldn't be true?
>

AFAIK, Bacula can generate messages which won't be combined in the mail
buffer file used to send emails.

There few differences which could be important:
> 1. You are using Bacula Enterprise and I am using Community version
>

The code handling messages is basically the same in both versions and AFAIK
for Community version above 9.6 the base code will be the same.

2. You are using S3 module and it might even create additional job,
>     I don't know because I newer used Bacula's S3 module.
>

It is the Storage Daemon plugin which does not create any additional jobs.
>From a technical point of view it creates a different archive device type,
so from a copy job perspective it is undistinguished from a tape drive.


> As for how they got combined, it's the same as with other combined
> messages.


Original job log is not handled by Storage Daemon. Storage Daemon generates
messages for the current job.


> In any case, Log table is not used for combining the messages from
> different Bacula daemons as it can be completely empty and one would
> still observe everything I was talking about in my previous posts.
>

The above is a perfect example that we are talking about two
totally distinctive things.
I'm talking about a log source for the original job log during copy. All
job messages combinations for email sending are handled by a dedicated file
buffer.
So, to combine two job logs in the file buffer for a copy job you have to
get an original job log from the log table of the catalog database and the
current job messages.
This combined log is then sent to the user and saved in the catalog as a
full clone job log.



> > I'm talking about logs from this job.
>
> But I am not talking about the original, normal Backup job.
>

The above is a perfect example that we are talking about two
totally distinctive things.


> That job which later gets copied by the Copy job has nothing with
> my post and its logs and report messages are fine.
>

I do not understand. How many emails do you get during a copy job?


>
> The problem is that when you run a Copy job which has to copy that
> normal Backup job, you get two e-mail messages instead of one.
> None of these two messages have nothing to do with the messages
> and logs produced while the original, normal Backup was running.
>

I do not understand. I'm always getting an original backup job log as a
part of the emails from copy jobs which are combined with the current job
copy logs.


>
> I hope it's now more clear what I am saying.
>
>
Unfortunate, no. :(


> >> The ID of the first Copy job which started the copy was 4999
> >> The ID that actually copied JobID 4997 from disk to tape was 5132
> >>
> >
> > which are combined with the above logs and saved as target logs.
>
> That's what I am saying, the logs related to 4999 and 5132 are not
> getting combined. Instead, you get two messages and the one related to
> the Job 5132 has the same Subject in the e-mail as the message sent
> for the normal, original Backup job.
>

During the copy job we have three jobid involved. The original backup job
log is combined with one of the remaining jobs as it becomes the new clone
- almost like an original.
I can verify it in my logs. I simply do not understand what you are getting
in yours and what in your opinion is a better option.


> But if one needs to specify one Copy job per normal Backup job,
> Bacula would send two e-mails per Copy job every time a Copy job
> is run.
>

Yes, because a copy job is a real two persons party. Everyone from these
two have its own runtime data to handle. I'm not judging if it is good or
not. I just made a fact.


> I don't think it's that complicated, it's just a matter of presenting
> it properly and understanding the issue.
>

I'm trying hard to understand the issue you expressed, but without any
luck. I do not get a point.
I suspect it is because we are both talking about two distinct things which
by accident are part of the copy job / cloning mechanism.


> I am observing this problem for years and if I get some time next
> year I will either open this issue or try to make a patch myself.
>

Good.


> I have also found a reproducible bug in Bacula which makes the Director
> segfault but it's not that critical.
>

You should report any segfaults as it is not good when the Director gets it.

best regards
-- 
Radosław Korzeniewski
rados...@korzeniewski.net
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users

Reply via email to