On 7/25/2013 9:58 AM, Yann Cézard wrote: > On 24/07/2013 16:26, Jeff Dickens wrote: >> Hi. I keep getting this error on one of my servers. >> >> A few words about the setup: >> >> The director is running under Ubuntu 12.04 LTS in a virtual machine >> on a very lightly loaded Xeon X3 with SSDs. >> >> The FD and SD are both running on a QNAP NAS box. I hand-built a >> bacula package for the QNAP. All are running 5.2.12. >> >> I have two identical QNAP NAS boxes, running the identical bacula >> package, and one works perfectly and has been extensively tested. >> The only difference is that one is connected to the director via a >> VPN (over Comcast Internet) and one is on the same LAN. >> >> The one on the remote QNAP will sometimes work for days and then >> sometimes only for a few hours. Restarting the SD will always make >> it work for a while. The log looks like this: >> >> 22-Jul 23:05 kiskadee-dir JobId 6445: No prior Full backup Job record >> found. >> 22-Jul 23:05 kiskadee-dir JobId 6445: No prior or suitable Full >> backup found in catalog. Doing FULL backup. >> 24-Jul 09:43 kiskadee-dir JobId 6445: Start Backup JobId 6445, >> Job=mill1-gene.2013-07-22_23.05.01_39 >> 24-Jul 09:43 kiskadee-dir JobId 6445: Purging oldest volume >> "mill1-gene-full-0117" >> 24-Jul 09:43 kiskadee-dir JobId 6445: 1 File on Volume >> "mill1-gene-full-0117" purged from catalog. >> 24-Jul 09:43 kiskadee-dir JobId 6445: There are no more Jobs >> associated with Volume "mill1-gene-full-0117". Marking it purged. >> 24-Jul 09:43 kiskadee-dir JobId 6445: All records pruned from Volume >> "mill1-gene-full-0117"; marking it "Purged" >> 24-Jul 09:43 kiskadee-dir JobId 6445: Using Device "FileStorage" >> 24-Jul 09:44 mill1-sd JobId 6445: Recycled volume >> "mill1-gene-full-0117" on device "FileStorage" (/share/bacula), all >> previous data lost. >> 24-Jul 09:43 kiskadee-dir JobId 6445: Max Volume jobs=1 exceeded. >> Marking Volume "mill1-gene-full-0117" as Used. >> 24-Jul 10:02 mill1-fd JobId 6445: Error: bsock.c:429 Write error >> sending 65809 bytes to Storage >> daemon:mill1.intranet.seamanpaper.com:9103 >> <http://mill1.intranet.seamanpaper.com:9103/>: ERR=Connection reset >> by peer >> 24-Jul 10:02 mill1-fd JobId 6445: Fatal error: backup.c:1200 Network >> send error to SD. ERR=Connection reset by peer >> 24-Jul 10:02 kiskadee-dir JobId 6445: Error: Director's comm line to >> SD dropped. >> 24-Jul 10:02 kiskadee-dir JobId 6445: Error: Bacula kiskadee-dir >> 5.2.12 (12Sep12): >> >> Thanks in advance for any light you can shed on this. >> >> -- >> *Jeff Dickens* >> IT Manager 978-632-1513 >> > Hi Jeff, > > You should read this : > http://wiki.bacula.org/doku.php?id=faq#my_backup_starts_but_dies_after_a_while_with_connection_reset_by_peer_error > Especially the part talking about keepalive (considering you are using > a VPN). > > Regards, > -- > Yann Cézard - administrateur systèmes serveurs > Centre de ressources informatiques -http://cri.univ-pau.fr > Université de Pau et des pays de l'Adour -http://www.univ-pau.fr
In addition to routers/switches/firewalls between the FD and SD, I have also seen the machine running FD cause this issue. For example, if a relatively weak client machine takes a long time to compress/encrypt a large file, it is possible that a network device driver with overly aggressive power management settings thinks there is no activity and powers down the PHY, even though there is an open socket. I don't know if that is a bug or a feature. I doubt that is the case with the QNAP box, but I have seen it with Windows and Mac clients, particularly when connected with WiFi devices. ------------------------------------------------------------------------------ See everything from the browser to the database with AppDynamics Get end-to-end visibility with application monitoring from AppDynamics Isolate bottlenecks and diagnose root cause in seconds. Start your free trial of AppDynamics Pro today! http://pubads.g.doubleclick.net/gampad/clk?id=48808831&iu=/4140/ostg.clktrk _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
