Hi, first of all, I am very happy for the existence of aufs, it fits nicely with the diskless networks I've setup at my customers. Unfortunately I have stumbled upon an error, which, even after long analysis, I could not find a solution for, so I humbly ask for help.
The Problem: Sometimes I loose all mounted filesystems on the client, except for the aufs mounted root. It is reproducible by triggering a tab filename completion in bash, it can mostly be prevented by doing an ls on the aufs mount before switching root at startup. Sometimes, with no clear evidence on what provoked it, all mounts are lost anyways. Using X11, OpenOffice, Firefox and Thunderbird is apparently usually safe, especially work on the command line seems to trigger it. The Error: Pressing ls /pro followed by the tab key (or /hom or /us or /var) All mounted filesystems are empty afterwards The Kludge: Putting ls /root > /dev/null in initramfs Filename completion is mostly safe in booted system The Setup: Debian Squeeze Server, custom kernel 3.6.10 (or 2.6.26 in another setup) with kernel-mode NFSv3 exports. Debian Squeeze clients, custom kernel 3.6.6 (3.6.10 in another setup) with aufs 3.6-20121112 as a module built with aufs3-standalone.git tree. Root mounting is done in a initramfs, custom script ( http://pastebin.com/2WB5dNXA ) /bin/mount -t tmpfs -o size=16M none /aufs/xino /bin/mount -t nfs -o rw,${NFSOPTS} ${SERVER}:/target/export/${HOSTNAME} /aufs/private /bin/mount -t nfs -o ro,${NFSOPTS} ${SERVER}:/target/export/clients /aufs/clients /bin/mount -t nfs -o ro,${NFSOPTS} ${SERVER}:/target/ /aufs/server /bin/mount -t aufs -o noplink,xino=/aufs/xino/root,dirs=/aufs/private=rw:/aufs/clients=ro+wh:/aufs/server=ro none ${TARGET} # HERE IS THE KLUDGE WORKING SOMEWHAT ls /target > /dev/null /bin/mount --move /aufs/xino ${TARGET}/aufs/xino /bin/mount --move /aufs/private ${TARGET}/aufs/private /bin/mount --move /aufs/clients ${TARGET}/aufs/clients /bin/mount --move /aufs/server ${TARGET}/aufs/server /bin/mount --move /sys ${TARGET}/sys /bin/mount --move /proc ${TARGET}/proc exec /bin/run-init ${TARGET} /sbin/init "$@" <${TARGET}/dev/console >${TARGET}/dev/console This appears similar to the observation of Jochen Rill http://sourceforge.net/mailarchive/message.php?msg_id=29515546 Trace of error ("ls /root" not done in initramfs) http://pastebin.com/gRVtxNfS Trace of no error ("ls /root" done in initramfs) http://pastebin.com/KmtBD32H Debugging option of aufs produces lots of output, mainly of the form: Jan 24 13:37:59 hti13 kernel: aufs au_do_h_d_reval:773:klogd[681]: DEBUG: b2 Jan 24 13:37:59 hti13 kernel: aufs h_d_revalidate:838:klogd[681]: DEBUG: b2, lo Full debug log while error is produced can be found here (large file, 46 MB): https://docs.google.com/file/d/0B0s7uiDVoB3OV09wNXZrM0Uyc2s/edit?usp=sharing If you need any further data I would be more than willing to provide it. Eibo Thieme /proc/mounts === rootfs / rootfs rw 0 0 none /sys sysfs rw,nosuid,nodev,noexec,relatime 0 0 none /proc proc rw,nosuid,nodev,noexec,relatime 0 0 none /aufs/xino tmpfs rw,relatime,size=16384k 0 0 192.168.0.1:/target/export/hti13/ /aufs/private nfs rw,relatime,vers=3,rsize=32768,wsize=32768,namlen=255,hard,nolock,proto=udp,timeo=11,retrans=10,sec=sys,mountaddr=192.168.0.1,mountvers=3,mountport=58461,mountproto=udp,local_lock=all,addr=192.168.0.1 0 0 192.168.0.1:/target/export/clients/ /aufs/clients nfs ro,relatime,vers=3,rsize=32768,wsize=32768,namlen=255,hard,nolock,proto=udp,timeo=11,retrans=10,sec=sys,mountaddr=192.168.0.1,mountvers=3,mountport=58461,mountproto=udp,local_lock=all,addr=192.168.0.1 0 0 192.168.0.1:/target/ /aufs/server nfs ro,relatime,vers=3,rsize=32768,wsize=32768,namlen=255,hard,nolock,proto=udp,timeo=11,retrans=10,sec=sys,mountaddr=192.168.0.1,mountvers=3,mountport=58461,mountproto=udp,local_lock=all,addr=192.168.0.1 0 0 none / aufs rw,relatime,si=6709df6d,noplink 0 0 tmpfs /lib/init/rw tmpfs rw,nosuid,relatime,mode=755 0 0 varrun /var/run tmpfs rw,nosuid,relatime,mode=755 0 0 varlock /var/lock tmpfs rw,nosuid,nodev,noexec,relatime 0 0 tmpfs /dev tmpfs rw,relatime,size=10240k,mode=755 0 0 tmpfs /dev/shm tmpfs rw,nosuid,nodev,relatime 0 0 devpts /dev/pts devpts rw,nosuid,noexec,relatime,gid=5,mode=620 0 0 192.168.0.1:/target/usr/ /usr nfs rw,noatime,vers=3,rsize=1048576,wsize=1048576,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=192.168.0.1,mountvers=3,mountport=58461,mountproto=udp,local_lock=none,addr=192.168.0.1 0 0 192.168.0.1:/target/var/ /aufs/server-var nfs ro,noatime,vers=3,rsize=1048576,wsize=1048576,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=192.168.0.1,mountvers=3,mountport=58461,mountproto=udp,local_lock=none,addr=192.168.0.1 0 0 192.168.0.1:/target/export/var/hti13/ /aufs/client-var nfs rw,noatime,vers=3,rsize=1048576,wsize=1048576,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=192.168.0.1,mountvers=3,mountport=58461,mountproto=udp,local_lock=none,addr=192.168.0.1 0 0 192.168.0.1:/target/export/tmp/hti13/ /tmp nfs rw,noatime,vers=3,rsize=1048576,wsize=1048576,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=192.168.0.1,mountvers=3,mountport=58461,mountproto=udp,local_lock=none,addr=192.168.0.1 0 0 192.168.0.1:/target/export/swaps/hti13/ /swap nfs rw,noatime,vers=3,rsize=1048576,wsize=1048576,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=192.168.0.1,mountvers=3,mountport=58461,mountproto=udp,local_lock=none,addr=192.168.0.1 0 0 192.168.0.1:/home/ /home nfs rw,noatime,vers=3,rsize=1048576,wsize=1048576,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=192.168.0.1,mountvers=3,mountport=58461,mountproto=udp,local_lock=none,addr=192.168.0.1 0 0 /sys/module/aufs/* === coresize 204507 initsize 0 initstate live refcnt 31 srcversion B884E6E7F16C3A1C018E993 taint O uevent cat: uevent: Keine Berechtigung version 3.6-20121112 parameters/brs 1 parameters/debug 0 sections/__bug_table 0xf89a040f sections/__ex_table 0xf89a2fc0 sections/__param 0xf899c1bc /sys/fs/aufs === config CONFIG_AUFS_FS=m CONFIG_AUFS_BRANCH_MAX_127=y CONFIG_AUFS_SBILIST=y CONFIG_AUFS_DEBUG=y si_6709df6d/br0 /aufs/private=rw si_6709df6d/br1 /aufs/clients=ro+wh si_6709df6d/br2 /aufs/server=ro si_6709df6d/xi_path /aufs/xino/root si_67260e6d/br0 /aufs/client-var=rw si_67260e6d/br1 /aufs/server-var=ro si_67260e6d/xi_path /aufs/xino/var client kernel config === => http://pastebin.com/DEd7NbzZ config.mk === # Kconfig # instead of setting 'n', leave it blank when you disable it. CONFIG_AUFS_BRANCH_MAX_127 = y CONFIG_AUFS_BRANCH_MAX_511 = CONFIG_AUFS_BRANCH_MAX_1023 = #CONFIG_AUFS_BRANCH_MAX_32767 = CONFIG_AUFS_SBILIST = y CONFIG_AUFS_HNOTIFY = y CONFIG_AUFS_HFSNOTIFY = CONFIG_AUFS_EXPORT = CONFIG_AUFS_RDU = CONFIG_AUFS_PROC_MAP = CONFIG_AUFS_SP_IATTR = CONFIG_AUFS_SHWH = CONFIG_AUFS_BR_RAMFS = CONFIG_AUFS_BR_FUSE = CONFIG_AUFS_BR_HFSPLUS = CONFIG_AUFS_DEBUG = y CONFIG_AUFS_MAGIC_SYSRQ = CONFIG_AUFS_BDEV_LOOP = CONFIG_AUFS_INO_T_64 = CONFIG_AUFS_POLL = ------------------------------------------------------------------------------ Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_feb
