I've been plagued by my backups often aborting with errors like this:
08-Jan 23:48 client-fd JobId 6: Error: bsock_tcp.c:405 Write error
sending 65536 bytes to Storage daemon:tape:8103: ERR=Permission denied
08-Jan 23:48 client-fd JobId 6: Fatal error: backup.c:1019 Network send error to SD. ERR=Permission denied
08-Jan 23:48 server-dir JobId 6: Error: Director's comm line to SD dropped.
08-Jan 23:48 server-dir JobId 6: Error: Bareos drtbrt-dir 16.2.7 (09Oct17):
I'm backing up 100s of TB to LTO-8 on an autoloader, this particular error happened 2 weeks into a 200TB backup with more than 150TB written. The logs on server and client show nothing indicating what the issue could be. It frustrates me, as why does bareos just give up and not wait and retry the send? The data is being spooled on the server, so from a write to tape issue, it's not streaming to the device, it's failing on the client to server send to the spool area. The Permission denied error reason seems strange, too. My only solution seems to be to create many small jobs to keep the backup times ~1 week, which is about 50TB. However I've also had this error occur after ~25TB backed up, it's not consistent. If there were network issues, I would think the OS logs would show problems. I realize my bareos version is rather old, but until I got to these very large filesystems, things were generally good. I'm running on FreeBSD currently, but will likely need to convert to Linux