Hi all: I'm trying to set up a new amanda server to back up my FreeNAS system for a customer. The FreeNAS system has a capacity of 125TB, and will probably be near there in a year or so; currently its about 93TB. I have a single LTO8 tape drive, currently running 8M formated tapes (9TB native).
I've set up a FreeBSD "Jail" (11.1) on the freenas server, mounted the storage volumes (read/write) inside the jail, and installed Amanda on the volume. This is an amanda client only. I have a dedicated PC (Intel i3) running CentOS 7 and Amanda from the repos (3.3.3) with the tape drive attached. This server only backs itself and the NAS up. I've broken up the DLEs as described on the wiki using include/exclude lists. I'm currently targeting a 11-12TB section of the NAS (9 DLEs). Last night's attempted run "kinda" succeeded. I have 3 DLEs that are under 10GB, and the rest are 500G-2.5T. All the estimates ran (I had to use checksize instead of client estimates), but the dumps on any of the larger DLEs failed. amstatus reports: amclient-tdrive:Vol1-2019-Session-A-D 0 failed: dumper: [request failed: error sending REQ: write error to: Broken pipe](too) amclient-tdrive:Vol1-2019-Session-E-J 0 failed: dumper: [request failed: error sending REQ: write error to: Broken pipe](too) amclient-tdrive:Vol1-2019-Session-K-M 0 failed: dumper: [request failed: error sending REQ: write error to: Broken pipe](too) amclient-tdrive:Vol1-2019-Session-N-Q 0 failed: dumper: [request failed: error sending REQ: write error to: Broken pipe](too) amclient-tdrive:Vol1-2019-Session-R-S 0 failed: dumper: [request failed: error sending REQ: write error to: Broken pipe](too) amclient-tdrive:Vol1-2019-Session-T-Z 0 failed: dumper: [request failed: error sending REQ: write error to: Broken pipe](too) amclient-tdrive:Vol1-2019-Session-rest 0 4g finished (22:35:03) amclient-tdrive:Vol1-General 0 failed: dumper: [request failed: error sending REQ: write error to: Broken pipe](too) amclient-tdrive:Vol2 0 failed: dumper: [request failed: error sending REQ: write error to: Broken pipe](too) localhost:/ 0 1g finished (22:36:07) localhost:/etc 0 0g finished (22:30:21) In the dumper logs on the server, I found this: Mon Apr 29 22:30:35 2019: thd-0x55b2cd007000: dumper: security_stream_seterr(0x55b2cd033bf0, write e rror to: Broken pipe) Mon Apr 29 22:30:35 2019: thd-0x55b2cd007000: dumper: security_seterror(handle=0x55b2cd02f520, drive r=0x7fcf6365e7e0 (BSDTCP) error=write error to: Broken pipe) Mon Apr 29 22:30:35 2019: thd-0x55b2cd007000: dumper: security_seterror(handle=0x55b2cd02f520, drive r=0x7fcf6365e7e0 (BSDTCP) error=error sending REQ: write error to: Broken pipe) Mon Apr 29 22:30:35 2019: thd-0x55b2cd007000: dumper: security_close(handle=0x55b2cd02f520, driver=0 x7fcf6365e7e0 (BSDTCP)) Mon Apr 29 22:30:35 2019: thd-0x55b2cd007000: dumper: security_stream_close(0x55b2cd033bf0) Mon Apr 29 22:30:35 2019: thd-0x55b2cd007000: dumper: security_stream_seterr(0x55b2cd033bf0, write e rror to: Broken pipe) I'm not finding much of use on the web for this error, or in the mailing list archives. Where do I go from here? Second, as this dump basically failed, but did manage to tape 3 DLEs, can I "void" this backup (remove it from index, mark it as not backed up, and the tape as reusable) and rerun on this same tape? Thanks! --Jim
