When putting /usr/portage in my /etc/exports, I got an error while nfs starting. Reproducible: Always Steps to Reproduce: 1. put /usr/portage 192.168.0.0/24(rw,no_root_squash,sync) in /etc/exports 2. /etc/init.d/nfs restart (or start) 3. Actual Results: $ sudo /etc/init.d/nfs restart * Stopping NFS mountd... [ ok ] * Stopping NFS daemon... [ ok ] * Stopping NFS statd... [ ok ] * Starting NFS statd... [ ok ] * Exporting NFS directories... /sbin/runscript.sh: line 529: 8223 Killed $exportfs -r 1>&2 * Error exporting NFS directories [ !! ] * Starting NFS daemon... [ ok ] * Starting NFS mountd... [ ok ] Expected Results: $ sudo /etc/init.d/nfs restart * Stopping NFS mountd... [ ok ] * Stopping NFS daemon... [ ok ] * Stopping NFS statd... [ ok ] * Starting NFS statd... [ ok ] * Exporting NFS directories... [ ok ] * Starting NFS daemon... [ ok ] * Starting NFS mountd... [ ok ] in /var/log/syslog: Jun 2 21:07:16 biproc rc-scripts: Error exporting NFS directories Jun 2 21:07:18 biproc kernel: svc: bad direction 65536, dropping request I have other exports in my file, but the problem only appear when /usr/portage is not commented :(
you shouldn't share /usr/portage via nfs anyway -- it screws with the metadata on other machines. best to have a local rsync mirror instead. It's quite alright, though, to share /usr/portage/disftiles.
Thanx for the quick answer. I didn't know there was some metadata in /usr/portage (why aren't they in /var/portage instead ?). Btw, is there a way to tell portage to remove unused archives (like apt) ? I don't want to do a find -mtime +30 because some packages like games-ftps/quake3-nsco need ~400 MB of files, and the first big archive (300 MB) is way too old (recent versions are just extension patches). I'll ask that on the forum. Since /usr/portage shouldn't be shared by nfs (for my purpose), I think we can close this ticket. Thanks again.
Nah, it's ok to put $PORTDIR on NFS, just makes things slower if you forgot to update the cache on the clients with `emerge --metadata`. Seems to be a kernel issue, noting portage related here.
It might be: I changed my kernel a couple of weeks ago from development-sources (2.6.5) to gentoo-development-sources (2.6.5-gentoo-r1). Since I use this nfs share to update a couple of laptops, I don't use it every day , and didn't detect the problem before. kind regards
is this ok to have : #uid = nobody #gid = nobody use chroot = no max connections = 20 pid file = /var/run/rsyncd.pid motd file = /etc/rsync/rsyncd.motd transfer logging = yes log format = %t %a %m %f %b syslog facility = local3 timeout = 300 [gentoo-x86-portage] #this entry is for compatibility path = /usr/portage comment = Gentoo Linux Portage tree [gentoo-portage] #modern versions of portage use this entry path = /usr/portage comment = Gentoo Linux Portage tree mirror exclude = distfiles in /etc/rsyncd/rsyncd.conf as a replacement ?
Hm, not a kernel bug I see...
Could this help? http://groups.google.com/groups?q=svc:+bad+direction&hl=en&lr=&ie=UTF-8&oe=UTF-8&scoring=d&selm=1rINI-3dU-23%40gated-at.bofh.it&rnum=1 From the few answers I have found out there, it does seem to be something to do with nfsd not liking some network traffic it's recieving. I was told to add: net.ipv4.ip_local_port_range = 49152 61000 to /etc/sysctl.conf But it didn't help in my case.