I have name service switch configured to use both files and LDAP for passwd and group database. Group lookup fails for users that are only in /etc/passwd, but not in LDAP: * User vglauche (only in LDAP) # id vglauche uid=606(vglauche) gid=100(users) groups=100(users),2005(recovery),2010(spm-kurs),2011(stn_op_mrt),2017(als) * User firebird (only in /etc/passwd) # id firebird uid=450(firebird) gid=450(firebird)id: failed to get groups for user `firebird': No such file or directory Reproducible: Always
Created attachment 143499 [details] emerge --info
Created attachment 143503 [details] nsswitch.conf
I can confirm this issue on sys-apps/coreutils-6.10-r1. It does not occur on sys-apps/coreutils-6.9-r1.
This bug has already been noticed upstream: http://thread.gmane.org/gmane.comp.gnu.coreutils.bugs/12375 Increasing severity - this bug prevents emerging any package that depends on id to check whether a certain user/group exists.
bug severity is irrelevant ... i'm not going to code any fixes, just wait for upstream to fix it and i'll merge those changes
*** Bug 210975 has been marked as a duplicate of this bug. ***
bug #210975 can reproduce this with: passwd: files winbind so it's anything with multi look ups. But I had a feeling it would be based on the e-mail on the ML.
added fixes from upstream to 6.10-r2
For me it happens also for users that are in both passwd and ldap. It also depends on what parameters are passed to "id" and sometimes, there is no error but still an incorrect result: (as root): > # id > uid=0(root) gid=0(root) groups=0(root),1(bin),2(daemon),3(sys),4(adm),6(disk), > 10(wheel),11(floppy),20(dialout),26(tape),27(video) > # id root > uid=0(root) gid=0(root) groups=0(root),1(bin),2(daemon),3(sys),4(adm) > # id <user> > uid=1001(<user>) gid=100(users)id: failed to get groups for user `<user>': No > such file or directory > # su - <user> > $ id > uid=1001(<user>) gid=100(users) groups=10(wheel),50(cvs),100(users) Otherwise, I confirm that sys-apps/coreutils-6.10-r2 fixes the problem. I don't think 6.10-r1 should have been marked stable with such a bug.
*** Bug 218221 has been marked as a duplicate of this bug. ***