Bug 745775
| Summary: | Unable to unmount autofs filesystems inside a container | |||
|---|---|---|---|---|
| Product: | Red Hat Enterprise Linux 6 | Reporter: | Daniel Berrangé <berrange> | |
| Component: | kernel | Assignee: | Ian Kent <ikent> | |
| Status: | CLOSED ERRATA | QA Contact: | yanfu,wang <yanwang> | |
| Severity: | unspecified | Docs Contact: | ||
| Priority: | unspecified | |||
| Version: | 6.3 | CC: | ikent, jmoyer, rwheeler, yanwang | |
| Target Milestone: | rc | |||
| Target Release: | --- | |||
| Hardware: | Unspecified | |||
| OS: | Unspecified | |||
| Whiteboard: | ||||
| Fixed In Version: | kernel-2.6.32-238.el6 | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | ||
| Clone Of: | ||||
| : | 920036 1048056 (view as bug list) | Environment: | ||
| Last Closed: | 2012-06-20 07:57:16 UTC | Type: | --- | |
| Regression: | --- | Mount Type: | --- | |
| Documentation: | --- | CRM: | ||
| Verified Versions: | Category: | --- | ||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | ||
| Cloudforms Team: | --- | Target Upstream Version: | ||
| Embargoed: | ||||
| Bug Depends On: | ||||
| Bug Blocks: | 920036, 1048056 | |||
| Attachments: | ||||
Since RHEL 6.2 External Beta has begun, and this bug remains unresolved, it has been rejected as it is not proposed as exception or blocker. Red Hat invites you to ask your support representative to propose this request, if appropriate and relevant, in the next release of Red Hat Enterprise Linux. Dan mentioned that this was fixed in the 3.1 kernel. Created attachment 558812 [details]
Patch - vfs: automount should ignore LOOKUP_FOLLOW
Created attachment 558813 [details]
Patch - vfs pathname lookup: Add LOOKUP_AUTOMOUNT flag
Created attachment 558814 [details]
Patch - VFS: Fix the remaining automounter semantics regressions
Created attachment 558815 [details]
Patch - vfs: remove LOOKUP_NO_AUTOMOUNT flag
The above patches are essentially the changes that went into kernel 3.1 that resolved the umount problem in Fedora. Let's give them a try here in RHEL please. This request was evaluated by Red Hat Product Management for inclusion in a Red Hat Enterprise Linux maintenance release. Product Management has requested further review of this request by Red Hat Engineering, for potential inclusion in a Red Hat Enterprise Linux Update release for currently deployed products. This request is not yet committed for inclusion in an Update release. Patch(es) available on kernel-2.6.32-238.el6 Created attachment 583470 [details]
Corrected problem demo program
I think this demo program has the corrections I described.
Please give it a try.
(In reply to comment #19) > Created attachment 583470 [details] > Corrected problem demo program > > I think this demo program has the corrections I described. > Please give it a try. Note that the test program is not intelligent at all so it can only check using simple mounts. So a hosts map with a tree of mounts isn't appropriate and can't be used. But what we want to test is if mounts inside the container can be umounted at all. Verified on kernel 2.6.32-269.el6: [root@sgi-xe320-01 ~]# rpm -qa|grep autofs autofs-5.0.5-54.el6.x86_64 [root@sgi-xe320-01 ~]# uname -a Linux sgi-xe320-01.rhts.eng.bos.redhat.com 2.6.32-269.el6.x86_64 #1 SMP Tue May 1 19:16:31 EDT 2012 x86_64 x86_64 x86_64 GNU/Linux [root@sgi-xe320-01 ~]# cat /etc/auto.master /net -hosts /- /etc/auto.marrow [root@sgi-xe320-01 ~]# cat /etc/auto.marrow /mnt/demo rhel6-nfs:/export/home [root@sgi-xe320-01 ~]# grep -E ' (/net|/mnt)' /proc/mounts -hosts /net autofs rw,relatime,fd=13,pgrp=26906,timeout=300,minproto=5,maxproto=5,indirect 0 0 /etc/auto.marrow /mnt/demo autofs rw,relatime,fd=19,pgrp=26906,timeout=300,minproto=5,maxproto=5,direct 0 0 [root@sgi-xe320-01 ~]# ./nsdemo /net We are the parent We are the container! Mount point 1 /net -hosts autofs rw,relatime,fd=13,pgrp=26906,timeout=300,minproto=5,maxproto=5,indirect Umount point 1 /net Done with umount [root@sgi-xe320-01 ~]# ls -l /net/rhel6-nfs total 0 dr-xr-xr-x. 3 root root 0 May 14 02:24 export [root@sgi-xe320-01 ~]# ./nsdemo /net/rhel6-nfs We are the parent We are the container! Done with umount [root@sgi-xe320-01 ~]# cat /proc/mounts -hosts /net autofs rw,relatime,fd=13,pgrp=27074,timeout=300,minproto=5,maxproto=5,indirect 0 0 /etc/auto.marrow /mnt/demo autofs rw,relatime,fd=19,pgrp=27074,timeout=300,minproto=5,maxproto=5,direct 0 0 -hosts /net/rhel6-nfs/export/home autofs rw,relatime,fd=13,pgrp=27074,timeout=300,minproto=5,maxproto=5,offset 0 0 [root@sgi-xe320-01 ~]# ./nsdemo /mnt/demo We are the parent We are the container! Mount point 1 /mnt/demo /etc/auto.marrow autofs rw,relatime,fd=19,pgrp=27074,timeout=300,minproto=5,maxproto=5,direct Umount point 1 /mnt/demo Done with umount [root@sgi-xe320-01 ~]# mount /dev/mapper/vg_sgixe32001-lv_root on / type ext4 (rw) proc on /proc type proc (rw) sysfs on /sys type sysfs (rw) devpts on /dev/pts type devpts (rw,gid=5,mode=620) tmpfs on /dev/shm type tmpfs (rw,rootcontext="system_u:object_r:tmpfs_t:s0") /dev/sda1 on /boot type ext4 (rw) /dev/mapper/vg_sgixe32001-lv_home on /home type ext4 (rw) none on /proc/sys/fs/binfmt_misc type binfmt_misc (rw) sunrpc on /var/lib/nfs/rpc_pipefs type rpc_pipefs (rw) nfsd on /proc/fs/nfsd type nfsd (rw) [root@sgi-xe320-01 ~]# ls -l /mnt/demo [root@sgi-xe320-01 ~]# cat /proc/mounts -hosts /net autofs rw,relatime,fd=13,pgrp=27074,timeout=300,minproto=5,maxproto=5,indirect 0 0 /etc/auto.marrow /mnt/demo autofs rw,relatime,fd=19,pgrp=27074,timeout=300,minproto=5,maxproto=5,direct 0 0 -hosts /net/rhel6-nfs/export/home autofs rw,relatime,fd=13,pgrp=27074,timeout=300,minproto=5,maxproto=5,offset 0 0 rhel6-nfs:/export/home /mnt/demo nfs4 rw,relatime,vers=4,rsize=524288,wsize=524288,namlen=255,hard,proto=tcp6,port=0,timeo=600,retrans=2,sec=sys,clientaddr=2620:52:0:1040:230:48ff:fe7f:2c8a,minorversion=0,local_lock=none,addr=2620:52:0:1040:221:5eff:fe95:20f4 0 0 [root@sgi-xe320-01 ~]# ./nsdemo /mnt/demo We are the parent We are the container! Mount point 1 /mnt/demo /etc/auto.marrow autofs rw,relatime,fd=19,pgrp=27074,timeout=300,minproto=5,maxproto=5,direct Mount point 2 /mnt/demo rhel6-nfs:/export/home nfs4 rw,relatime,vers=4,rsize=524288,wsize=524288,namlen=255,hard,proto=tcp6,port=0,timeo=600,retrans=2,sec=sys,clientaddr=2620:52:0:1040:230:48ff:fe7f:2c8a,minorversion=0,local_lock=none,addr=2620:52:0:1040:221:5eff:fe95:20f4 Umount point 2 /mnt/demo Umount point 1 /mnt/demo Done with umount Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. http://rhn.redhat.com/errata/RHSA-2012-0862.html |
Created attachment 527958 [details] Demo to illustrate umount problems inside a container Description of problem: One of the first things that is required when setting up a container with a private root filesystem, is to unmount all filesystems inherited from the host OS. For most filesystems this works without trouble, but for autofs this is not the case. For demonstration, consider an autofs map with one direct mount and one indirect mount: # cat /etc/auto.master /net -hosts /- /etc/auto.marrow # cat /etc/auto.marrow /mnt/demo marrow.example.com:/var/lib/libvirt/images When autofs initially starts these mount points are visible # grep -E ' (/net|/mnt)' /proc/mounts -hosts /net autofs rw,relatime,fd=6,pgrp=2938,timeout=300,minproto=5,maxproto=5,indirect 0 0 /etc/auto.marrow /mnt/demo autofs rw,relatime,fd=12,pgrp=2938,timeout=300,minproto=5,maxproto=5,direct 0 0 The attached demo program creates a container and attempts to unmount the requested filesystem inside the container. First trying to unmount '/net' fails: # ./autofsdemo /net We are the parent We are the container! Found mount point 1 /net -hosts autofs rw,relatime,fd=6,pgrp=2938,timeout=300,minproto=5,maxproto=5,indirect Umount point 1 /net Umount point 0 /net Could not umount /net: Invalid argument If we then trigger a mount within that location # ls /net/marrow.example.com bin boot cgroup dev etc home lib lost+found media mnt opt proc root sbin selinux srv sys tmp usr var And now try to unmount that volume inside the container # ./autofsdemo /net/marrow.example.com We are the parent We are the container! Found mount point 1 /net/marrow.example.com marrow.example.com:/ nfs rw,nosuid,nodev,relatime,vers=3,rsize=524288,wsize=524288,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=10.1.1.2,mountvers=3,mountport=35386,mountproto=udp,local_lock=none,addr=10.1.1.2 Umount point 1 /net/marrow.example.com Could not umount /net/marrow.example.com: Device or resource busy Next up consider the other autofs location which is not triggered yet: # ./autofsdemo /mnt/demo We are the parent We are the container! Found mount point 1 /mnt/demo /etc/auto.marrow autofs rw,relatime,fd=12,pgrp=2938,timeout=300,minproto=5,maxproto=5,direct Umount point 1 /mnt/demo Could not umount /mnt/demo: Too many levels of symbolic links Some odd has now happened though. The container has managed to trigger the mounting of this location in the host OS: # grep /mnt /proc/mounts /etc/auto.marrow /mnt/demo autofs rw,relatime,fd=12,pgrp=2938,timeout=300,minproto=5,maxproto=5,direct 0 0 marrow.example.com:/var/lib/libvirt/images /mnt/demo nfs rw,relatime,vers=3,rsize=524288,wsize=524288,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=10.1.1.2,mountvers=3,mountport=35386,mountproto=udp,local_lock=none,addr=10.1.1.2 0 0 Re-run the demo again, shows we can at least unmount the NFS volume this time, but still can't unmount the autofs point # ./autofsdemo /mnt/demo We are the parent We are the container! Found mount point 1 /mnt/demo /etc/auto.marrow autofs rw,relatime,fd=12,pgrp=2938,timeout=300,minproto=5,maxproto=5,direct Found mount point 2 /mnt/demo marrow.example.com:/var/lib/libvirt/images nfs rw,relatime,vers=3,rsize=524288,wsize=524288,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=10.1.1.2,mountvers=3,mountport=35386,mountproto=udp,local_lock=none,addr=10.1.1.2 Umount point 2 /mnt/demo Umount point 1 /mnt/demo Could not umount /mnt/demo: Too many levels of symbolic links So there are multiple bugs here 1. You can't unmount autofs filesystems of type 'indirect' - Invalid argument 2. You can't unmount NFS filesystems underneath an 'indirect' autofs mount - Device or resource busy 3. You can't unmount autofs filesystems of type 'direct' - Too many levels of symbolic links 4. Attempting to umount direct mounts, actually triggers mounting of the NFS volume in the host OS None of this would be fatal, if it were possible to use 'umount2($path, MNT_DETACH)' to hide the mount from the container, but that also fails. Version-Release number of selected component (if applicable): kernel 2.6.32-207.el6.x86_64 autofs-5.0.5-23.el6.x86_64 How reproducible: Always Steps to Reproduce: 1. See above