[Bug 813641] New: NFSv4 idmapper broken: rpc.idmapd: nss_getpwnam: name '1000' does not map into domain 'bla.test'
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c0 Summary: NFSv4 idmapper broken: rpc.idmapd: nss_getpwnam: name '1000' does not map into domain 'bla.test' Classification: openSUSE Product: openSUSE 11.4 Version: Final Platform: x86-64 OS/Version: openSUSE 11.4 Status: NEW Severity: Normal Priority: P5 - None Component: Kernel AssignedTo: kernel-maintainers@forge.provo.novell.com ReportedBy: gleixner@bib-bvb.de QAContact: qa-bugs@suse.de Found By: --- Blocker: --- User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:19.0) Gecko/20100101 Firefox/19.0 I have a NFSv4 Server using Opensuse 12.2 and two NFSv4 Clients, one Opensuse 12.2 and another 11.4 (evergreen). All 3 machines have kerberos/ldap for authentication and can resolve ldap entries correctly using getent passwd 1000 or getent passwd flo The clients use the same automounter maps (via ldap) to mount homes. The maps force the clients to use NFSv4. All have exactly the same /etc/idmapd.conf with the domain set to "bla.test". The opensuse 12.2 NFS client gets the correct ownership of the mounted homes, while the opensuse 11.4 client gets nobody/nobody. Running rpc.idmapd -f i get: rpc.idmapd: libnfsidmap: using domain: bla.test rpc.idmapd: libnfsidmap: loaded plugin /usr/lib64/libnfsidmap/nsswitch.so for method nsswitch rpc.idmapd: Expiration time is 600 seconds. rpc.idmapd: nfsdopenone: Opening /proc/net/rpc/nfs4.nametoid/channel failed: errno 2 (No such file or directory) rpc.idmapd: New client: 4c2 rpc.idmapd: Opened /var/lib/nfs/rpc_pipefs/nfs/clnt4c2/idmap rpc.idmapd: New client: 4c3 rpc.idmapd: nfs4_name_to_uid: calling nsswitch->name_to_uid rpc.idmapd: nss_getpwnam: name '0' domain 'bla.test': resulting localname '(null)' rpc.idmapd: nss_getpwnam: name '0' does not map into domain 'bla.test' rpc.idmapd: nfs4_name_to_uid: nsswitch->name_to_uid returned -22 rpc.idmapd: nfs4_name_to_uid: final return value is -22 rpc.idmapd: Client 4c2: (user) name "0" -> id "65534" rpc.idmapd: nfs4_name_to_gid: calling nsswitch->name_to_gid rpc.idmapd: nfs4_name_to_gid: nsswitch->name_to_gid returned -22 rpc.idmapd: nfs4_name_to_gid: final return value is -22 rpc.idmapd: Client 4c2: (group) name "0" -> id "65533" rpc.idmapd: New client: 4c4 rpc.idmapd: nfs4_name_to_uid: calling nsswitch->name_to_uid rpc.idmapd: nss_getpwnam: name '1000' domain 'bla.test': resulting localname '(null)' rpc.idmapd: nss_getpwnam: name '1000' does not map into domain 'bla.test' rpc.idmapd: nfs4_name_to_uid: nsswitch->name_to_uid returned -22 rpc.idmapd: nfs4_name_to_uid: final return value is -22 rpc.idmapd: Client 4c2: (user) name "1000" -> id "65534" rpc.idmapd: nfs4_name_to_gid: calling nsswitch->name_to_gid rpc.idmapd: nfs4_name_to_gid: nsswitch->name_to_gid returned -22 rpc.idmapd: nfs4_name_to_gid: final return value is -22 rpc.idmapd: Client 4c2: (group) name "1000" -> id "65533" I guess nfsidmap package is buggy? Reproducible: Always Steps to Reproduce: 1. 2. 3. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c1 Rob Verduijn <rob.verduijn@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |rob.verduijn@gmail.com Component|Kernel |Kernel Product|openSUSE 11.4 |openSUSE 12.3 --- Comment #1 from Rob Verduijn <rob.verduijn@gmail.com> 2013-07-09 08:13:35 UTC --- Hi, There seems to be a problem with pulse audio when I use a home drive mounted on an nfs4 share. I get this in the message log : pulseaudio[2907]: [pulseaudio] core-util.c: Failed to create secure directory (/home/rob/.config/pulse): Invalid argument pulseaudio[2903]: [pulseaudio] main.c: Daemon startup failed. the directory /home/rob/.config/pulse already exists and is writable by the user rob Pulse works fine when I mount the directory with nfs3 Rob ps I changed the suse version to 12.3 since it's definitely not about 11.4 -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c2 Jeff Mahoney <jeffm@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |jeffm@suse.com AssignedTo|kernel-maintainers@forge.pr |nfbrown@suse.com |ovo.novell.com | --- Comment #2 from Jeff Mahoney <jeffm@suse.com> 2013-07-15 14:12:19 EDT --- Comment #1 is unrelated to the original issue. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c3 Neil Brown <nfbrown@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #3 from Neil Brown <nfbrown@suse.com> 2013-07-15 21:25:48 UTC --- I think you are saying that opensuse 12.3 works but opensuse 11.4 works. Is that correct? In that case, I recommend using opensuse 12.3. 11.4 is no longer supported. If I misunderstood, please re-open the bug. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c4 Rob Verduijn <rob.verduijn@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED | --- Comment #4 from Rob Verduijn <rob.verduijn@gmail.com> 2013-07-16 07:01:03 UTC --- This is still a problem in openSUSE 12.3 It is still biting me and forcing me to use NFS3 while I would prefer NFS4 Thus I reopened Regards Rob -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c Rob Verduijn <rob.verduijn@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- OS/Version|openSUSE 11.4 |openSUSE 12.3 -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c5 --- Comment #5 from Rob Verduijn <rob.verduijn@gmail.com> 2013-07-16 07:01:54 UTC --- I noticed the version of opensuse was still on 11.3, adjusted now its on 12.3 Rob -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c6 Neil Brown <nfbrown@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |NEEDINFO InfoProvider| |gleixner@bib-bvb.de --- Comment #6 from Neil Brown <nfbrown@suse.com> 2013-07-16 07:20:02 UTC --- Sorry, but I don't understand. You said:
The opensuse 12.2 NFS client gets the correct ownership of the mounted homes, while the opensuse 11.4 client gets nobody/nobody.
So presumably 11.4 is broken and 12.2 is fixed. Now you say that 12.3 is broken. Can you explain exactly how 12.3 is broken. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c7 Rob Verduijn <rob.verduijn@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |REOPENED InfoProvider|gleixner@bib-bvb.de | --- Comment #7 from Rob Verduijn <rob.verduijn@gmail.com> 2013-07-16 08:44:50 UTC --- When I mount my homedrive using nfs4 my pulse audio fails, when I mount it using nfs3 pulse audio works. Looking into the logs I found [pulseaudio] core-util.c: Failed to create secure directory (/home/rob/.config/pulse): Invalid argument [pulseaudio] main.c: Daemon startup failed. This dissapears when I mount the homedrive with nfs3 Furthermore [pulseaudio] core-util.c: Failed to create secure directory (/home/rob/.config/pulse): Invalid argument [pulseaudio] main.c: Daemon startup failed. Which is weird since the directory exists and is writable by the user. This also starts working when mounting with nfs3 Also my problem sounds an awfull lot like this bug: https://bugzilla.novell.com/show_bug.cgi?id=785803 Which refers to : https://bugzilla.redhat.com/show_bug.cgi?id=829362 Where they talk about the idmapper. Which brings me to this ticket that I found using the search in bugzilla, querrying nfs and pulse. Rob -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c8 Neil Brown <nfbrown@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |NEEDINFO InfoProvider| |rob.verduijn@gmail.com --- Comment #8 from Neil Brown <nfbrown@suse.com> 2013-07-23 06:04:33 UTC --- Oh, I see. This is totally different bug that you thought might be related because it was somehow connected to idmapd?? If you have different symptoms you really should open a different bug. Let us join them if they turn out to have the same root cause. But seeing the original bug was against an no-longer supported release we may as well continue you yours here. Can you get a tcpdump trace of the NFSv4 network traffic? tcpdump -s 0 -i $INTERFACE -w /tmp/nfs.pcap port 2049 and then try starting pulseaudio, or whatever you do to trigger the error. Then attach the 'nfs.pcap' file. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c9 --- Comment #9 from Rob Verduijn <rob.verduijn@gmail.com> 2013-07-28 10:11:19 UTC --- Created an attachment (id=549820) --> (http://bugzilla.novell.com/attachment.cgi?id=549820) tcpdump of the nfs traffic -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c10 Rob Verduijn <rob.verduijn@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |REOPENED InfoProvider|rob.verduijn@gmail.com | --- Comment #10 from Rob Verduijn <rob.verduijn@gmail.com> 2013-07-28 10:15:31 UTC --- Hello, Sorry for the delay. I've uploaded the requested pcap (see previous message) The error is easy to reproduce. Mount /home on a nfs4 export. Create an ordinary user. Log in as that user. And watch the message log, pulse will start generating the error messages shortly after the user logged in. Some sidenotes that probably don't matter but just in case. I use a bridge configured by yast for kvm. (br0) My system partition is btrfs. Cheers Rob -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c11 Neil Brown <nfbrown@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |NEEDINFO InfoProvider| |rob.verduijn@gmail.com --- Comment #11 from Neil Brown <nfbrown@suse.com> 2013-07-29 04:28:25 UTC --- What kernel are you running on the server? It seems to be a pre-2.6.38 kernel. Upgrading that probably wouldn't hurt, but might not make a big difference. That kernel is returning the wrong error code for an unknown username. When pulseaudio tried to "chown" the newly created directory is request the ownership to be set to "rob@localdomain" and the group to "users@localdomain". The server doesn't like this for some reason and reports NFS4ERR_BADNAME (which should be NFS4ERR_BADOWNER). So you need to find out why idmapper isn't working properly on the server. What distro? What kernel? Is idmapper running? Are there errors in the logs on the server? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c12 --- Comment #12 from Rob Verduijn <rob.verduijn@gmail.com> 2013-07-29 08:15:07 UTC --- Damn, The idmapper isn't running on the destination server. I've missed that bloody obvious error, I really need a vacation. I should have checked the nfs4 server. The kernel release is indeed old 2.6.32. It's a synology nas, and upgrading the kernel is a bit of a pain to make a serious understatement. They lag behind with their software even though they do regular updates. Anyway, gonna take this up with their support team. Thank you for your help and sorry for wasting your time. Rob -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c13 Rob Verduijn <rob.verduijn@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |REOPENED InfoProvider|rob.verduijn@gmail.com | --- Comment #13 from Rob Verduijn <rob.verduijn@gmail.com> 2013-07-29 08:15:49 UTC --- forgot the needinfo flag -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c14 Randall Smith <rbsmith@sandia.gov> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |rbsmith@sandia.gov --- Comment #14 from Randall Smith <rbsmith@sandia.gov> 2013-07-29 14:48:23 UTC --- We are experiencing the same problem as the original issue (and rpc.idmapd is running on the server). We just upgraded our NFSv4 servers from OpenSuSE 11.2 to OpenSuSE 12.3 and now uids and gids on our older 11.x OpenSuSE and Ubunt clients map to nobody:nobody. This behavior essentially broke our entire network. The version differences in the affected packages are minor (nfsidmap v0.23 versus v0.24 and nfs-client v1.2.3 versus v1.2.7) so I wouldn't expect backward compatibility to be broken. None of the 12.x release notes make mention of this. NFSv4 is such a pervasive and important service, I think the backward compatibility issue deserves a bit more attention. For a quick work-around, we built an OpenSuSE 11.4 KVM NFS server to hold us over until we can understand the problem and find a fix. It's not always possible to upgrade all of the affected clients. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c15 Neil Brown <nfbrown@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |NEEDINFO InfoProvider| |rbsmith@sandia.gov --- Comment #15 from Neil Brown <nfbrown@suse.com> 2013-07-29 23:47:19 UTC --- I would need a tcpdump network trace (see comment #8 above) and idmap logs. Run rpc.idmapd with -fvvvv and collect the output. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c16 --- Comment #16 from Randall Smith <rbsmith@sandia.gov> 2013-08-01 23:04:33 UTC --- Just to clarify, our NFSv4 id mapping issue is identical to flo gleixner's original submission; the rpc.idmapd verbose output looks identical. We also employ ldap so our /etc/nsswitch.conf directs lookups to ldap. Does your request for traces and logs suggest that NFSv4 id mapping should work with older NFSv4 clients like OpenSuse 11.x and Ubuntu 11.x? For example, is compatibility expected between nfs-utils-1.2.7-2.10.1 and nfs-utils-1.2.3-11.16.1? Here are a few observations: After killing rpc.idmapd on both our OpenSuSE 12.3 NFSv4 server and our OpenSuSE 12.1 NFSv4 client, we find that ls -l shows that uid and gid are still correctly mapped to names on the client. However for the case of an OpenSuSE 11.4 NFSv4 server and an OpenSuSE11.4 NFSv4 client, just killing rpc.idmapd on the client, ls -l shows the uid and gid unmapped as 4294967294. With rpc.idmapd running on the client and server, and after several exchanges with both 12.1 and 11.4 clients, the contents of /proc/net/rpc/nfs4.nametoid/content on the OpenSuSE 12.3 server has only the header "#domain type name [id]". The same goes for nfs4.idtoname/content. A network trace on the 11.4 client reveals that the V4 REPLY to a V4 LOOKUP contains (correct) numeric UID and GIDs for the fattr4_owner and fattr4_owner_group attributes. Would you still like me to send in a network trace and the rpc.idmapd outputs? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c17 Randall Smith <rbsmith@sandia.gov> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |REOPENED InfoProvider|rbsmith@sandia.gov | --- Comment #17 from Randall Smith <rbsmith@sandia.gov> 2013-08-05 11:40:50 UTC --- The NFSv4 id mapping compatibility between an OpenSuSE 12.3 server and OpenSuSE 11.4 client is resolved in comment #1 of https://bugzilla.novell.com/show_bug.cgi?id=810408 -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=813641 https://bugzilla.novell.com/show_bug.cgi?id=813641#c18 Neil Brown <nfbrown@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED --- Comment #18 from Neil Brown <nfbrown@suse.com> 2013-08-06 06:49:08 UTC --- Only 4 months ago - you'd think I would remember... So that fixes it for you I take it? Great! Please re-open if I misunderstood. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠novell.com