This is a discussion on amanda and xinetd within the Gentoo Linux Support forums, part of the Unix Operating Systems category; --> I'm having trouble getting amanda working with gentoo. I've never used either amanda or xinetd before, and I'm new ...
| |||||||
| FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
| ||||
| I'm having trouble getting amanda working with gentoo. I've never used either amanda or xinetd before, and I'm new to gentoo as well, so I may not be interpreting the symptoms properly, but I get the impression that xinetd is not invoking /usr/libexec/amandad when I think it's supposed to. Specifically, executing su amanda -c "/usr/sbin/amcheck daily" gives these results: Amanda Tape Server Host Check ----------------------------- amcheck-server: slot 6: date X label daily6 (new tape) NOTE: skipping tape-writable test Tape daily6 label ok Server check took 0.384 seconds Amanda Backup Client Hosts Check -------------------------------- WARNING: lotus: selfcheck request timed out. Host down? Client check: 1 host checked in 30.030 seconds, 1 problem found (brought to you by Amanda 2.4.4p3) /var/log/messages contains Apr 17 18:38:29 lotus su(pam_unix)[12607]: session opened for user amanda by john(uid=0) Apr 17 18:38:29 lotus xinetd[12616]: FAIL: amanda address from=192.168.1.101 Apr 17 18:38:29 lotus xinetd[9729]: START: amanda pid=12616 from=192.168.1.101 Apr 17 18:38:29 lotus xinetd[9729]: EXIT: amanda status=0 pid=12616 duration=0(sec) Apr 17 18:38:39 lotus xinetd[12632]: FAIL: amanda address from=192.168.1.101 Apr 17 18:38:39 lotus xinetd[9729]: START: amanda pid=12632 from=192.168.1.101 Apr 17 18:38:39 lotus xinetd[9729]: EXIT: amanda status=0 pid=12632 duration=0(sec) Apr 17 18:38:49 lotus xinetd[12633]: FAIL: amanda address from=192.168.1.101 Apr 17 18:38:49 lotus xinetd[9729]: START: amanda pid=12633 from=192.168.1.101 Apr 17 18:38:49 lotus xinetd[9729]: EXIT: amanda status=0 pid=12633 duration=0(sec) Apr 17 18:38:59 lotus su(pam_unix)[12607]: session closed for user amanda which tells me that xinetd didn't like something about trying to start amandad. "lotus" is the name of the machine on which amanda is installed and on which I'm executing all the commands. /etc/xinetd.d/amanda contains the recommended configuration: service amanda { socket_type = dgram protocol = udp wait = yes user = amanda group = amanda groups = yes server = /usr/libexec/amandad only_from = localhost disable = no } lsof | grep amanda gives this output: xinetd 12820 root 5u IPv4 19164 UDP *:amanda xinetd 12820 root 6u IPv4 19165 TCP *:amandaidx (LISTEN) and netstat -a | grep -i amanda gives tcp 0 0 *:amandaidx *:* LISTEN udp 0 0 *:amanda *:* I'm running a fairly recent version of gentoo, built from stage 1 in the last 2 weeks. I'd appreciate any pointers about how to fix what amcheck is complaining about. John Blinka |
| |||
| So anyway, it was like, 00:57 CEST Apr 18 2005, you know? Oh, and, yeah, John Blinka was all like, "Dude, > Apr 17 18:38:39 lotus xinetd[12632]: FAIL: amanda address from=192.168.1.101 > only_from = localhost Try removing the restriction to localhost as a source address. hth. -- Time flies like an arrow, fruit flies like a banana. Perth ---> * 12:56:26 up 176 days, 22 min, 10 users, load average: 0.03, 0.02, 0.00 Linux 2.6.9 x86_64 GNU/Linux Registered Linux user #261729 |
| |||
| Johan Lindquist wrote: > So anyway, it was like, 00:57 CEST Apr 18 2005, you know? Oh, and, yeah, > John Blinka was all like, "Dude, > > >>Apr 17 18:38:39 lotus xinetd[12632]: FAIL: amanda address from=192.168.1.101 > > >> only_from = localhost > > > Try removing the restriction to localhost as a source address. > > hth. > I've done that. It generates a different error: ERROR: lotus: [access as amanda not allowed from amanda@lotus] open of /var/spool/amanda/.amandahosts failed At the moment, I'm only trying to run backups from localhost to localhost, so the restriction should not be a problem. Besides, I'm now of the opinion that xinetd is not the source of my troubles, since nmap -sU localhost gives Starting nmap 3.75 ( http://www.insecure.org/nmap/ ) at 2005-04-18 11:06 EDT Interesting ports on localhost (127.0.0.1): (The 1476 ports scanned but not shown below are in state: closed) PORT STATE SERVICE 631/udp open|filtered unknown 10080/udp open|filtered amanda indicating that xinetd is listening at port 10080. But /usr/libexec/amandad, which should be invoked after an appropriate communication at port 10080, shows no access via the ls -lu command after doing the amcheck. So, xinetd is listening, but it doesn't fire up amandad. Looking at the ports being accessed via iftop during amcheck, I see only traffic on port 674, which seems very strange. What does 674 have to do with amanda? John Blinka |
| |||
| So anyway, it was like, 17:21 CEST Apr 18 2005, you know? Oh, and, yeah, John Blinka was all like, "Dude, > Johan Lindquist wrote: >> So anyway, it was like, 00:57 CEST Apr 18 2005, you know? Oh, and, yeah, >> John Blinka was all like, "Dude, >>>Apr 17 18:38:39 lotus xinetd[12632]: FAIL: amanda address from=192.168.1.101 >>> only_from = localhost >> >> Try removing the restriction to localhost as a source address. > > I've done that. It generates a different error: > > ERROR: lotus: [access as amanda not allowed from amanda@lotus] open > of /var/spool/amanda/.amandahosts failed That's quite logical then. > At the moment, I'm only trying to run backups from localhost to > localhost, The log snippet you provided seems to say that you're failing to do this, since the attempts aren't coming from localhost but from (I'm presuming) the actual ip address of the host. I haven't (yet) finished my own amanda setup, so I'm not sure I can help you further. Just make sure you're actually using localhost as the server, and not the hostname. -- Time flies like an arrow, fruit flies like a banana. Perth ---> * 19:28:06 up 176 days, 6:54, 9 users, load average: 0.06, 0.03, 0.01 Linux 2.6.9 x86_64 GNU/Linux Registered Linux user #261729 |
| |||
| Johan Lindquist wrote: > > That's quite logical then. > > >>At the moment, I'm only trying to run backups from localhost to >>localhost, > > > The log snippet you provided seems to say that you're failing to do > this, since the attempts aren't coming from localhost but from (I'm > presuming) the actual ip address of the host. Yes, they are. > > I haven't (yet) finished my own amanda setup, so I'm not sure I can > help you further. Just make sure you're actually using localhost as > the server, and not the hostname. > I think you may have touched on the difficulty. While doing something else, I discovered that the amanda emerge populated /etc/csh.env and /etc/profile.env with all sorts of environmental variables: printenv | grep AMANDA AMANDA_SERVER_INDEX=lotus AMANDA_USER_GROUPS=amanda AMANDA_PORTS_UDP= AMANDA_USER_HOMEDIR=/var/spool/amanda AMANDA_GROUP_NAME=amanda AMANDA_USER_UID=87 AMANDA_CONFIG_NAME=DailySet1 AMANDA_PORTS_TCP= AMANDA_PORTS_BOTH= AMANDA_SERVER=lotus AMANDA_TMPDIR=/var/spool/amanda/tmp AMANDA_DBMODE=db AMANDA_PORTS= AMANDA_TAR_LISTDIR=/var/spool/amanda/tar-lists AMANDA_USER_SH=/bin/false AMANDA_GROUP_GID=87 AMANDA_USER_NAME=amanda AMANDA_SERVER_TAPE=lotus These values appear to be in conflict with at least some of /etc/xinetd.d/amanda and /etc/amanda/... Now there's something to explore. John Blinka |
| ||||
| So anyway, it was like, 00:48 CEST Apr 19 2005, you know? Oh, and, yeah, John Blinka was all like, "Dude, [..] > populated /etc/csh.env and /etc/profile.env with all sorts of > environmental variables: The file you'll probably have to edit is /etc/env.d/97amanda. When you've made the appropriate changes, don't forget to run 'env-update'. -- Time flies like an arrow, fruit flies like a banana. Perth ---> * 08:34:31 up 176 days, 20:00, 11 users, load average: 0.00, 0.00, 0.00 Linux 2.6.9 x86_64 GNU/Linux Registered Linux user #261729 |