[Lustre-discuss] selinux and lustre

[Leif Bergman]

Hello,

I've run into an odd error when attempting to mount the lustre MDT.

I've found that when I set the selinux policy to 'disabled' it works
fine, but if it is set to 'permissive' I get the error below:


Dec 31 15:32:31 ops03 kernel: Lustre: OBD class driver,
http://www.lustre.org/
Dec 31 15:32:31 ops03 kernel:         Lustre Version: 1.6.6
Dec 31 15:32:31 ops03 kernel:         Build Version:
1.6.6-19691231190000-PRISTINE-.cache.OLDRPMS.20080826155503.linux-smp-2.6.18-92.1.10.el5_lustre.1.6.6.i686-2.6.18-92.1.10.el5_lustre.1.6.6smp
Dec 31 15:32:31 ops03 kernel: Lustre: Added LNI 192.168.8.25 at tcp
[8/256]
Dec 31 15:32:31 ops03 kernel: Lustre: Accept secure, port 988
Dec 31 15:32:32 ops03 kernel: Lustre: Lustre Client File System;
http://www.lustre.org/
Dec 31 15:32:32 ops03 kernel: kjournald starting.  Commit interval 5
seconds
Dec 31 15:32:32 ops03 kernel: LDISKFS FS on dm-6, internal journal
Dec 31 15:32:32 ops03 kernel: LDISKFS-fs: recovery complete.
Dec 31 15:32:32 ops03 kernel: LDISKFS-fs: mounted filesystem with
ordered data mode.
Dec 31 15:32:32 ops03 kernel: kjournald starting.  Commit interval 5
seconds
Dec 31 15:32:32 ops03 kernel: LDISKFS FS on dm-6, internal journal
Dec 31 15:32:32 ops03 kernel: LDISKFS-fs: mounted filesystem with
ordered data mode.
Dec 31 15:32:32 ops03 kernel: Lustre: MGS MGS started
Dec 31 15:32:32 ops03 kernel: Lustre: Enabling user_xattr
Dec 31 15:32:32 ops03 kernel: Lustre: MDT testfs-MDT0000 now serving
dev (testfs-MDT0000/65218f61-7818-9990-4fe1-5edd077c1e3d) with
recovery enabled
Dec 31 15:32:32 ops03 kernel: Lustre:
6075:0:(lproc_mds.c:273:lprocfs_wr_group_upcall()) testfs-MDT0000:
group upcall set to /usr/sbin/l_getgroups
Dec 31 15:32:32 ops03 kernel: Lustre: testfs-MDT0000.mdt: set
parameter group_upcall=/usr/sbin/l_getgroups
Dec 31 15:32:32 ops03 kernel: Lustre: Server testfs-MDT0000 on device
/dev/VolGroup00/LogVolLustre has started
Dec 31 15:32:32 ops03 kernel: SELinux: (dev lustre, type lustre) has
no xattr support
Dec 31 15:32:32 ops03 kernel: Lustre: Failing over testfs-MDT0000
Dec 31 15:32:32 ops03 kernel: Lustre: *** setting obd testfs-MDT0000
device 'unknown-block(253,6)' read-only ***
Dec 31 15:32:32 ops03 kernel: Turning device dm-6 (0xfd00006)
read-only
Dec 31 15:32:32 ops03 kernel: Lustre: testfs-MDT0000: shutting down
for failover; client state will be preserved.
Dec 31 15:32:32 ops03 kernel: Lustre: MDT testfs-MDT0000 has stopped.
Dec 31 15:32:32 ops03 kernel: LustreError:
5822:0:(ldlm_request.c:996:ldlm_cli_cancel_req()) Got rc -108 from
cancel RPC: canceling anyway
Dec 31 15:32:32 ops03 kernel: LustreError:
5822:0:(ldlm_request.c:1605:ldlm_cli_cancel_list())
ldlm_cli_cancel_list: -108
Dec 31 15:32:32 ops03 kernel: Lustre: MGS has stopped.
Dec 31 15:32:32 ops03 kernel: Removing read-only on dm-6 (0xfd00006)
Dec 31 15:32:32 ops03 kernel: Lustre: server umount testfs-MDT0000
complete



Anybody run in to this? It seems odd that a 'permissive' setting should
cause it to break.

Regards,

Leif Bergman
Renesys