[pkg-bacula-devel] Bug#1009231: Write errors since upgrade to 9.6.7-4

[Klaus Ethgen]

Hi Sven,

Am Fr den 15. Apr 2022 um 10:52 schrieb Sven Hartge:
> > Since upgrade from 9.6.7-3 to 9.6.7-4 I get many of the following errors
> > in the log. However, the backup seems to work.
> > 
> > JobId 0: Security Alert: bsock.c:380 Write error sending 4 bytes to client:127.0.0.1:57050: ERR=Broken pipe
> > JobId 0: Security Alert: bsock.c:380 Write error sending 4 bytes to client:127.0.0.1:57074: ERR=Broken pipe
> > ...
> > JobId 0: Security Alert: bsock.c:380 Write error sending 4 bytes to client:127.0.0.1:57134: ERR=Connection reset by peer
> > JobId 0: Security Alert: bsock.c:380 Write error sending 4 bytes to client:127.0.0.1:57126: ERR=Connection reset by peer
> > ...
> 
> Hello Klaus,
> 
> I have not been able to reproduce this nor am I seeing this on any of my
> systems.
> 
> Please check that the version of bacula-sd matches the one from
> bacula-director and that bacula-fd is of no higher version than the Director
> and the SD.

That is the case:
   ~> dpkg -l bacula\*
   Gewünscht=Unbekannt/Installieren/R=Entfernen/P=Vollständig Löschen/Halten
   | Status=Nicht/Installiert/Config/U=Entpackt/halb konFiguriert/
	    Halb installiert/Trigger erWartet/Trigger anhängig
   |/ Fehler?=(kein)/R=Neuinstallation notwendig (Status, Fehler: GROSS=schlecht)
   ||/ Name                     Version      Architektur  Beschreibung
   +++-========================-============-============-======================================================
   ii  bacula-bscan             9.6.7-4      amd64        network backup service - bscan tool
   ii  bacula-common            9.6.7-4      amd64        network backup service - common support files
   ii  bacula-common-sqlite3    9.6.7-4      amd64        network backup service - SQLite v3 common files
   ii  bacula-console           9.6.7-4      amd64        network backup service - text console
   ii  bacula-console-qt        9.6.7-4      amd64        network backup service - Bacula Administration Tool
   ii  bacula-director          9.6.7-4      amd64        network backup service - Director daemon
   ii  bacula-director-sqlite3  9.6.7-4      all          network backup service - SQLite 3 storage for Director
   ii  bacula-doc               9.6.7-1      all          Documentation for Bacula
   ii  bacula-fd                9.6.7-4      amd64        network backup service - file daemon
   ii  bacula-sd                9.6.7-4      amd64        network backup service - storage daemon

Funny fact: I also have systems with version 9.4.2 FD doing the backup
to this system and they do not trigger my problem over the network. Only
the localhost connections from the very same system does.

> Also seeing "Connection reset by peer" while connecting via localhost is
> very suspicious, IMHO. Is the FD dying during the backup?

I thought the same.

And no, the fd is not dying and the backup is completed without other
troubles.

But maybe I have some hints. About a half a year ago I added a host
living in untrusted network, so I changed the configuration to be mixed
clear communication over trusted network and encrypted for the rest.
Therefor, I have the following in director config:

   Client xxx
   {
      ...
      TLS Enable = yes
      TLS Require = yes
      TLS CA Certificate File = /etc/ssl/private/bacula-ca.pem
      TLS Certificate = /etc/ssl/private/bacula-dir.cert
      TLS Key = /etc/ssl/private/bacula-dir.key
   }

All other TLS settings are off.

I also have a nagios bacula monitor running every 5 minutes.
(/usr/lib/nagios/plugins/check_bacula) But this seems to stay green all
the time.

> What do you system logs and dmesg show during the exact time this happens?
> 
> > ii  init-system-helpers      1.62devuan1
> 
> This looks like a mixed Debian/Devuan system. Can you make sure to reproduce
> the bug in a clean Debian system, to rule out anything being affected by
> changes made for Devuan.

No, I completely migrated to devuan some years ago due to the
systemd-debakel. I do not have any debian system left.

Gruß
   Klaus
-- 
Klaus Ethgen                                       http://www.ethgen.ch/
pub  4096R/4E20AF1C 2011-05-16            Klaus Ethgen <Klaus at Ethgen.ch>
Fingerprint: 85D4 CA42 952C 949B 1753  62B3 79D0 B06F 4E20 AF1C
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 688 bytes
Desc: not available
URL: <http://alioth-lists.debian.net/pipermail/pkg-bacula-devel/attachments/20220415/ce700a00/attachment-0002.sig>