Hi, On Fri, 2009-07-31 at 15:26 -0300, Rogério Brito wrote: > [...] > > My USB-disk (not flash) has two partitions: 1. vfat 2. ext3. Both > > partitions are recognized and have no errors. > > Right. It is strange, but I had a problem quite similar to this one with > some udev versions, but I can't reproduce the problem anymore (I'm using > sid here, with, perhaps, some updated udev). > > Can you still reproduce things with udev + usbmount from sid?
I am using sid already and there're no new versions of udev available other than 0.141-1 ... > > > When the disk is connected, it is recognized and usbmount mounts them > > with correct FS types, but both partitions on the exact same location > > (/media/usb0): > > > > -------- excerpt from /var/log/syslog -------- > > Jul 1 00:55:06 sokrates usbmount[21916]: mountpoint /media/usb0 is > > available for /dev/sdb2 > > Jul 1 00:55:06 sokrates usbmount[21920]: mountpoint /media/usb0 is > > available for /dev/sdb1 > > -------- end excerpt from /var/log/syslog -------- > > > > It seems that usbmount is in a race when checking the available > > mountpoint and find it both times available ... it's just a matter of > > luck which partition is mounted first. This might lead to data loss, > > since a standard user might not recognize that there is such problem. > > Yes, it seems that there is some race condition somewhere and I will try > to check with lockfile-progs what can be done. There're also no new version of lockfile-progs other than 0.1.13 ... I tried with usbmount 0.0.14-1 which I found in svn and the disk is mounted fine: -------- excerpt from /var/log/syslog -------- Aug 5 16:12:51 sokrates kernel: [1807358.676659] sdb: sdb1 sdb2 Aug 5 16:12:51 sokrates kernel: [1807358.723532] sd 16:0:0:0: [sdb] Attached SCSI disk Aug 5 16:12:51 sokrates kernel: [1807358.724131] sd 16:0:0:0: Attached scsi generic sg1 type 0 Aug 5 16:12:52 sokrates usbmount[5130]: executing command: mount -tvfat -osync,noexec,nodev,noatime /dev/sdb1 /media/usb0 Aug 5 16:12:52 sokrates usbmount[5130]: executing command: run-parts /etc/usbmount/mount.d Aug 5 16:12:57 sokrates usbmount[5129]: executing command: mount -text3 -osync,noexec,nodev,noatime /dev/sdb2 /media/usb1 Aug 5 16:12:57 sokrates kernel: [1807365.256657] kjournald starting. Commit interval 5 seconds Aug 5 16:12:57 sokrates kernel: [1807365.256657] EXT3 FS on sdb2, internal journal Aug 5 16:12:57 sokrates kernel: [1807365.256657] EXT3-fs: mounted filesystem with ordered data mode. Aug 5 16:12:57 sokrates usbmount[5129]: executing command: run-parts /etc/usbmount/mount.d -------- end excerpt from /var/log/syslog -------- I believe that 0.0.14-1 seem to work fine, since it doesn't try to lock (or doesn't show it); with this there is a difference of +/- 5 seconds between the two mounts, while 0.0.17 does it at the "same" second. Unfortunately I don't have enough time left (yet) to go deeper in this and check it; for now I'll pin usbmount to 0.0.14-1 ... > > Thanks for your bugreport, Rogério Brito. > > > P.S.: Just out of curiosity, which program do you use to report bugs? I > see it is different from what my reportbug program uses to list the > packages dependencies... reportbug-ng 1.5 from sid. Regards -- hgb
signature.asc
Description: This is a digitally signed message part