SAMBA Share Service are too slow

Hi to all.
Files services are really too slow. It take more 15 mns for 512 Mo !
All network connection are clear, no wifi, wire cable.
No dmesg error.

Smbd error:
an Invalid argument error. This can happen when using 64 bit lock offsets
on 32 bit NFS mounted file systems.
[2020/04/20 19:30:52.017536, 0] …/source3/locking/posix.c:272(posix_fcntl_getlock)
posix_fcntl_getlock: WARNING: lock request at offset 35389440, length 262144 returned
an Invalid argument error. This can happen when using 64 bit lock offsets
on 32 bit NFS mounted file systems.
[2020/04/20 19:30:54.320857, 0] …/source3/locking/posix.c:272(posix_fcntl_getlock)
posix_fcntl_getlock: WARNING: lock request at offset 35651584, length 524288 returned
an Invalid argument error. This can happen when using 64 bit lock offsets
on 32 bit NFS mounted file systems.

I saw some settings on smb.conf
How to changing the smb.conf for some tcp setting tests ?

Any one get same ?

I do have the problem with NFS in file manager and since 9.2.2 file copy/move the speed just drops lower and lower and I have to cancel the operation.

I have had this issue on the N2 with low speeds since I got it, but it always completed the copying, I raised the issue in the beginning but we didn’t come to a solution and the test on my network was stellar.

1 Like

I have the same problem with a 922x-h device on playing video. When I fiirst install the 9.2.2 everything works well. But after several days. It doesn’t work. 4k uhd became slow. There is tips on the screen that my source is too slow to read continuously. But I have tested in win10 that read and write speed is all 1000 Mbps speed.
One more,when I switch to the android tv kodi , everything is good. All uhd source can play without any problems.

I tested the new night release from 15 April.
The issue seems to be ok. It’s more speedy, and working.
I was transfered 6,75GB in aurounding 10/12 mns from wire Ethernet cable thru Netgear Orb Satellite to Wire cable.
And all is complete !

No error during the connection and files transfer.
But apparently at the startup machine:
[2020/04/24 10:37:22.321262, 0] …/source3/locking/posix.c:272(posix_fcntl_getlock)
posix_fcntl_getlock: WARNING: lock request at offset 30932992, length 294092 returned
an Invalid argument error. This can happen when using 64 bit lock offsets
on 32 bit NFS mounted file systems.

I done a fresh restart and connect back to samba share.
same error on logs. But it’s seem to be stable about the share connection.

[2019/09/03 09:27:22.545243, 0] …/source3/smbd/server.c:1754(main)
smbd version 4.9.18 started.
Copyright Andrew Tridgell and the Samba Team 1992-2018
[2019/09/03 09:27:22.615739, 0] …/source3/passdb/pdb_smbpasswd.c:249(startsmbfilepwent)
startsmbfilepwent_internal: file /run/samba/smbpasswd did not exist. File successfully created.
[2019/09/03 09:27:22.641253, 0] …/lib/util/become_daemon.c:136(daemon_ready)
daemon_ready: daemon ‘smbd’ finished starting up and ready to serve connections
[2020/04/24 10:44:24.637240, 0] …/source3/locking/posix.c:272(posix_fcntl_getlock)
posix_fcntl_getlock: WARNING: lock request at offset 0, length 6148 returned
an Invalid argument error. This can happen when using 64 bit lock offsets
on 32 bit NFS mounted file systems.
[2020/04/24 10:44:24.639704, 0] …/source3/locking/posix.c:272(posix_fcntl_getlock)
posix_fcntl_getlock: WARNING: lock request at offset 0, length 6148 returned
an Invalid argument error. This can happen when using 64 bit lock offsets
on 32 bit NFS mounted file systems.

CoreELEC uses a 64 bit kernel and 32 bit userspace, so this sounds like an expected message.

I have had a more serious problem but with the same error messages so I though I should be posting here.

When I’m playing files from samba share (samba client is CoreELEC Matrix, samba server is CoreELEC Leila), playback stops after 2 GB or so, and resuming from that point also fails (player stucks in loading circle).

Journalctl of samba server shows the same messages. Despite it saying “Warning”, these are errors and file is unreadable after 2GB point.

Solution is adding strict locking = no to [global] section of smb.conf.

Since you say this problem is expected in CoreELEC, I propose strict locking to be disabled by default in CoreELEC’s default samba configuration.

Have a nice day.

Can you post more details about server/client system, filesystem and so on?

I use only Samba and read files from a Windows NAS with NTFS and never had such issues.

Server: Samba server of CoreELEC 9.2.8 with default config
Shares are the ones automatically created by CoreELEC of external hard drives. Drives are formatted as EXT3.

File is 4,5 GB mkv file. But it doesn’t have to be a media file, trying to read offset after 2GB of any file should be triggering the same error.

Samba error:

../source3/locking/posix.c:272(posix_fcntl_getlock)
  posix_fcntl_getlock: WARNING: lock request at offset 34078720, length 65536 returned
  an Invalid argument error. This can happen when using 64 bit lock offsets
  on 32 bit NFS mounted file systems.

Kodi error:

Prepare - Start position set to last stopped position: 2639000
CVideoPlayerVideo - video decoder returned error
Skipped 87 duplicate messages..
CVideoPlayerVideo - video decoder returned error
Skipped 207 duplicate messages..

This is not a Kodi error, even a simple file copying via Thunar etc. fails after the same offset.