Finally, it is fixed by "chmod 777 /". http://www.opensolaris.org/jive/thread.jspa?forumID=13&threadID=69913 I checked libavl.so.1 and it seems OK. rpcbind's not running seems caused by datalink-management failure . But thanks anyway.
-Harry > * Harry Fu <harry.fu at sun.com> [2008-08-20 05:59]: > > Recently I keep encountering a same problem that my > solaris became > > unusable by unknown reason. It seems that > datalink-management and > > rpc-bind services stopped randomly and can not be > started again. Hence > > I can not boot to multi-user milestone. I have no > idea about how to > > get this fixed or how to find out the root cause. > Attached are the > > error messages and logs I got but it doesn't give > much useful > > information to me. > > From your attachment, I would investigate two > things: > 1. What are the permissions on /lib/libavl.so.1? > Is it a valid ELF > file? (That is, why can't the linker read it?) > > On a recent system, > $ ls -l /lib/libavl.so.1 > ~/project/pkg/repo/pkg-clone > xr-x 1 root bin 14724 Jun 19 10:43 /lib/libavl.so.1* > $ elfdump /lib/libavl.so.1 > /dev/null > $ > > (elfdump will print an "invalid file type" > message to stderr for a > non-ELF file. Success rules out some, but not > all, corruption > scenarios...) > You probably need to truss rpcbind to determine > what's causing it > to fail. > Stephen > > -- > sch at sun.com http://blogs.sun.com/sch/ > _______________________________________________ > smf-discuss mailing list > smf-discuss at opensolaris.org This message posted from opensolaris.org
