Bruno, thanks for reply. I have already checked the firewall - there is nothing set when looking either using iptables nor nftables and all interfaces involved are on the same subnet.
Are you suggesting adding the heartbeat or ensure it is not added... unsure what you mean. I normally don't have it enabled and have not needed it so far. I can try passive client, certainly. It would be irritating to need it but hey ho. The SD is running directly on (Debian-12 based) proxmox (to avoid issues getting access to the tape drives), while the director and the client are both in VMs. I am aware proxmox does some network stuff... might this be getting in the way? I don't understand how, if so, but it is the only thing I can think of. Ruth On Monday, 17 February 2025 at 10:57:22 UTC Bruno Friedmann (bruno-at-bareos) wrote: > Can be sure at 100% > but you certainly have to investigate in connection drop, > Maybe you need to check if you can setup `heartbeat interval` > https://docs.bareos.org/TasksAndConcepts/NetworkSetup.html#heartbeat-interval-tcp-keepalive, > > set client as passive > https://docs.bareos.org/TasksAndConcepts/NetworkSetup.html#passive-clients > > check any intrusive firewall that might cut idle tcp connection too > agressivly. > > Hope this help you a bit. > > On Monday, 17 February 2025 at 11:23:08 UTC+1 [email protected] wrote: > >> Folks, >> >> I am repeatedly getting an issue with multiple clients where the backup >> starts ok and gets to the point of writing to tape and then fails, as shown >> in the log attached. Snippet below: >> >> 16-Feb 18:07 greyarea-bareos-dir JobId 15775: Connected Client: >> greyarea-dns-fd at greyarea-dns.cam.ivimey.org:9102, encryption: >> TLS_CHACHA20_POLY1305_SHA256 TLSv1.3 >> ... >> 16-Feb 18:07 greyarea-dns-fd JobId 15775: Connected Storage daemon at >> helva.cam.ivimey.org:9103, encryption: TLS_CHACHA20_POLY1305_SHA256 >> TLSv1.3 >> ... >> 16-Feb 18:07 helva-sd JobId 15775: Spooling data ... >> 16-Feb 18:07 greyarea-dns-fd JobId 15775: Fatal error: TLS read/write >> failure.: ERR=error:0A000126:SSL routines::unexpected eof while reading >> 16-Feb 18:07 greyarea-bareos-dir JobId 15775: Fatal error: >> lib/tls_openssl_private.cc:357 TLS read/write failure.: >> ERR=error:0A000126:SSL routines::unexpected eof while reading >> 16-Feb 18:07 greyarea-bareos-dir JobId 15775: Fatal error: Director's >> comm line to SD dropped. >> >> It has been happening for a while, with multiple clients and jobs, and >> has me utterly baffled especially as that there are jobs for some clients >> that successfully complete. Can anyone help? >> >> The director, storage and client are all running apt-installed community >> builds tagged 24.0.0~pre1546.c16dbcf30-8, which was from 14th Dec 2024. The >> OS is Ubuntu 24.04 'noble' running on amd64 cpus, and all but 'storage' are >> qemu VMs. >> >> Possibly relevant is that I also have some clients which the director >> seems unable to contact, even though 'ss -lt' shows the process is >> listening, 'tcpdump' shows connection requests reach the client system on >> the right port, and 'ping' works fine in both directions. an 'strace' of >> the bareos-fd process shows it is not even attempting to respond to the >> connection request from the director. This may or may not be related... ^o^. >> >> >> >> -- You received this message because you are subscribed to the Google Groups "bareos-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion visit https://groups.google.com/d/msgid/bareos-users/3b260f4b-b7f1-44b5-afd6-9efff6722716n%40googlegroups.com.
