Florian Fainelli
2022-Aug-10 02:34 UTC
[Samba] macOS monterey unable to back up to Samba 4.14.12
On 8/9/2022 7:33 PM, Christopher Snowhill wrote:> > >> On Aug 9, 2022, at 7:08 PM, Florian Fainelli <f.fainelli at gmail.com> wrote: >> >> ? >> >> On 8/8/2022 8:57 PM, Christopher Snowhill wrote: >>>>> On Aug 8, 2022, at 8:22 PM, Florian Fainelli via samba <samba at lists.samba.org> wrote: >>>> >>>> ?Hi all, >>>> >>>> I have been trying different configurations to get one of my macOS Monterey machine to back up to a samba4 installation on a Raspberry Pi 4B using OpenWrt. Since this is an OpenWrt installation the package is heavily stripped to its minimum so here is the list of vfs modules being installed: >>>> >>>> ls /usr/lib/samba/vfs/ >>>> cap.so commit.so default_quota.so fake_perms.so fruit.so offline.so recycle.so shell_snap.so virusfilter.so worm.so >>>> catia.so crossrename.so dirsort.so fileid.so io_uring.so readonly.so shadow_copy2.so streams_xattr.so widelinks.so xattr_tdb.so >>>> >>>> The error reported by the Mac client is that a sparsebundle file gets created but ultimately the backup preparation does not go through and shows "The backup disk image could not be created". >>>> >>>> During the short time where the .sparsebundle folder is created and visible (it gets deleted upon failure), the permissions all look similar to those of the other working backup. >>>> >>>> The Mac does complain about not being able to read a number of files under the .sparsebundle folder that I can create just fine using Finder or the OpenWrt device itself, so it sounds like as if the files were attempted to be created with attributes that are unsupported and we failed to read them back. >>>> >>>> The same samba 4 installation works great on older Macs though. Underlying filesystem is ext4. >>>> >>>> My relevant smb.conf section for that share is below: >>>> >>>> [TimeMachine] >>>> path = /mnt/backup/TimeMachine >>>> create mask = 0666 >>>> directory mask = 0777 >>>> read only = no >>>> guest ok = yes >>>> guest only = yes >>>> fruit:encoding = native >>>> fruit:metadata = stream >>>> fruit:veto_appledouble = no >>>> fruit:nfs_aces = no >>>> fruit:posix_rename = yes >>>> fruit:wipe_intentionally_left_blank_rfork = yes >>>> inherit permissions = yes >>>> fruit:time machine = yes >>>> vfs objects = catia fruit streams_xattr io_uring >>>> >>>> Thanks for reading me! >>>> -- >>>> Florian >>>> >>>> -- >>>> To unsubscribe from this list go to the following URL and read the >>>> instructions: https://lists.samba.org/mailman/options/samba >>> I found at least that the error was occurring because Samba is crashing and restarting. I intended to acquire access to the bug tracker to file a report, but they never replied to my request. >> >> That does not appear to be the case for me. >> -- >> Florian > > Then you have something else to file a report on, it seems. I'm building debug tevent and samba packages to get a better backtrace on the crash.Is the bug tracker the best way to get the attention of someone knowledgeable about what could be happening, or is the mailing list equally appropriate? -- Florian
Christopher Snowhill
2022-Aug-10 02:55 UTC
[Samba] macOS monterey unable to back up to Samba 4.14.12
> On Aug 9, 2022, at 7:34 PM, Florian Fainelli <f.fainelli at gmail.com> wrote: > > ? > > On 8/9/2022 7:33 PM, Christopher Snowhill wrote: >>>> On Aug 9, 2022, at 7:08 PM, Florian Fainelli <f.fainelli at gmail.com> wrote: >>> >>> ? >>> >>> On 8/8/2022 8:57 PM, Christopher Snowhill wrote: >>>>>> On Aug 8, 2022, at 8:22 PM, Florian Fainelli via samba <samba at lists.samba.org> wrote: >>>>> >>>>> ?Hi all, >>>>> >>>>> I have been trying different configurations to get one of my macOS Monterey machine to back up to a samba4 installation on a Raspberry Pi 4B using OpenWrt. Since this is an OpenWrt installation the package is heavily stripped to its minimum so here is the list of vfs modules being installed: >>>>> >>>>> ls /usr/lib/samba/vfs/ >>>>> cap.so commit.so default_quota.so fake_perms.so fruit.so offline.so recycle.so shell_snap.so virusfilter.so worm.so >>>>> catia.so crossrename.so dirsort.so fileid.so io_uring.so readonly.so shadow_copy2.so streams_xattr.so widelinks.so xattr_tdb.so >>>>> >>>>> The error reported by the Mac client is that a sparsebundle file gets created but ultimately the backup preparation does not go through and shows "The backup disk image could not be created". >>>>> >>>>> During the short time where the .sparsebundle folder is created and visible (it gets deleted upon failure), the permissions all look similar to those of the other working backup. >>>>> >>>>> The Mac does complain about not being able to read a number of files under the .sparsebundle folder that I can create just fine using Finder or the OpenWrt device itself, so it sounds like as if the files were attempted to be created with attributes that are unsupported and we failed to read them back. >>>>> >>>>> The same samba 4 installation works great on older Macs though. Underlying filesystem is ext4. >>>>> >>>>> My relevant smb.conf section for that share is below: >>>>> >>>>> [TimeMachine] >>>>> path = /mnt/backup/TimeMachine >>>>> create mask = 0666 >>>>> directory mask = 0777 >>>>> read only = no >>>>> guest ok = yes >>>>> guest only = yes >>>>> fruit:encoding = native >>>>> fruit:metadata = stream >>>>> fruit:veto_appledouble = no >>>>> fruit:nfs_aces = no >>>>> fruit:posix_rename = yes >>>>> fruit:wipe_intentionally_left_blank_rfork = yes >>>>> inherit permissions = yes >>>>> fruit:time machine = yes >>>>> vfs objects = catia fruit streams_xattr io_uring >>>>> >>>>> Thanks for reading me! >>>>> -- >>>>> Florian >>>>> >>>>> -- >>>>> To unsubscribe from this list go to the following URL and read the >>>>> instructions: https://lists.samba.org/mailman/options/samba >>>> I found at least that the error was occurring because Samba is crashing and restarting. I intended to acquire access to the bug tracker to file a report, but they never replied to my request. >>> >>> That does not appear to be the case for me. >>> -- >>> Florian >> Then you have something else to file a report on, it seems. I'm building debug tevent and samba packages to get a better backtrace on the crash. > > Is the bug tracker the best way to get the attention of someone knowledgeable about what could be happening, or is the mailing list equally appropriate? > -- > FlorianThe bug tracker may be the best place to file a report with logs, but the mailing list may be the best place to get the attention needed to get registered on the tracker, if you've already tried and failed with the request address.