similar to: implement of persistent handle in SMB2.1

hi Michael, Sorry for taking so long to reply, I just get back on vacation. Your advice is very helpful. Our purpose is no IO interrupted with Window 7 cifs client accessing the shared folder when virtual IP taken over by another node. SMB3.0 or newer protocol do meet that demand, but Window 7 cifs client only supports SMB2.1 protocol and previous versions. So we need implement a

On 2016-09-28 at 12:50 +0000, Weidong wrote: > hi Michael： Hi Weidong, Thank you for your message! > We want to implement the ‘persistent handle’ function under > Windows 7(SMB 2.1), which has only local ‘durable handle’. Hmm. As you say: Windows 7 / SMB 2.1 does _not_ support presistent handles. This requires SMB 3.0 or newer. Windows 7 supports durable handles, which is already

Hi list We recently upgraded our fileservers from Centos supplied 4.2.10 to Sernet 4.4.6, and then our DCs from 3.6.x to 4.4.6. It seems that since then we've had problems with locks not being obeyed on all nodes - they only seem to work when a second client opens a file on the same node as the first client. For example, when a user opens an Excel file I will see something like this

On 27/08/15 22:20, Robert Moskowitz wrote: > > > On 08/27/2015 05:10 PM, Rowland Penny wrote: >> On 27/08/15 22:00, Robert Moskowitz wrote: >>> Ah, LDAP is included within Samba, I find. Don't install provided >>> one... >>> >>> I suppose I will have to find what schemas, particularly if the bind >>> dlz schema is included?

Hello, I'm having a permissions problem I just can't quite figure out. I can not access any directory that is owned by my userid if the permissions are 700. If I change them to 770, with my default group on the directory, then I can access the directory. This was working fine with Fedora 19 for several years, and it broke when I upgraded to Fedora 22, and still broken with F23. my

On 3/3/2016 2:02 PM, Marc Muehlfeld wrote: > It would help if you would provide some information about your system: > - Samba version - smb.conf - "ls -ld" of the directory - etc. Regards, > Marc Hello, version: samba-4.3.4-1.fc23.x86_64 ls -ld: ls -ld /home/blaster drwxrwx--- 76 blaster users 4096 Mar 3 14:15 /home/blaster testparm Load smb config files from

Hi Everyone, My doubts are for Samba4.1 performance related enhancements (SMB2.1 / SMB3.0). My experimental samba server hardware is running 4.1 version. Want to improve robocopy WRITE throughput from a two samba client machines (win7 or win8 or server2012 x86 servers). Someone please advice me how can I use SMB2.1 or SMB3.0 capabilities to improve WRITE throughput. I tried samba4.1

Thanks very much, Volker. I want to synchronize three tdb files(listed as below) which is relevant to the opened file: 1) brlock.tdb 2) locking.tdb 3) smbXsrv_open_global.tdb These three tdb files are used to realize the durable file handle feature. As what you have said, the performance will be affected if storing these three files in the shared storage. So can I realize the synchronization

Hi all Thanks for your attention. CTDB manages NICs with event scripts. Virtual IP attaches on a NIC as a secondary IP. When noticing the NIC down, CTDB will pull it up. Does any one know how to disable the action? I do not want CTDB pull any NIC up. Does the command 'ctdb disablescript 10.interface' has any unexpected effect? Thanks.

Hi folks I face a problem about the data synchronization when using the ctdb. When one ctdb node fails, other ctdb nodes may not recover the data of the failed node. I think it may be due to the fact that ctdb stores its tdb files in the local storage. So, is it fine to store the tdb files in the shared directory of the backend storage? When one node fails, other nodes may recover the data from

If eth0 holds virtual IP as its secondary IP and I bring down eth0 by command "ifdown eth0", CTDB will automatically pull eth0 up. Sometime it goes wrong. The virtual IP becomes primary IP, and origin primary IP turns into a secondary IP or just missing. If the origin primary IP is missing, I have to run "ifup eth0" though CTDB has pulled eth0 up. CTDB uses “ip link set eth0

Hi all http://linux.die.net/man/1/ctdbd. In this page, it says “This directory is local for each node and should not be stored on the shared cluster filesystem”. But why? Thanks for reply. --dbdir=<directory> This is the directory on local storage where ctdbd keeps the local copy of the TDB databases. This directory is local for each node and should not be stored on the shared cluster

I've just upgraded my DM from samba 4.5 to 4.8, using luois repos (also, debian jessie -> stretch). I'm experimenting some troubles on some shares, seems that both guest access and 'machine account' access does not work. The share is rather simple: [wpkg] browseable = No comment = WPKG Automated Software Deploying System force create mode = 0664 force directory mode =

Hi, We are running a samba fileserver, access controlled using posix acl (right 770, with users/groups on the filesystem level. Therefore samba shares look like this: [share] path = /srv/academic read only = no writable = yes create mask = 0770 directory mask = 0770 Now certain users complain that they cannot access certain folders, but looking at the folders from the linux fileystem, their

On 20/10/16 14:10, Amitay Isaacs wrote: > On Thu, Oct 20, 2016 at 8:09 PM, Alex Crow via samba > <samba at lists.samba.org <mailto:samba at lists.samba.org>> wrote: > > Hi list > > We recently upgraded our fileservers from Centos supplied 4.2.10 > to Sernet 4.4.6, and then our DCs from 3.6.x to 4.4.6. > > It seems that since then we've

Thank you very much for reply. Our test team insists on simulating network failure by using ifdown instead of just unpluging the cable. I have to modify the event script 10.interface. When noticing a NIC failure, CTDB just marks the NIC down and do not try to pull it up. It seems to work fine. -----邮件原件----- 发件人: Martin Schwenke [mailto:martin at meltin.net] 发送时间: 2016年9月18日 18:49 收件人: Weidong

On 27.12.2015 20:48, Marc Muehlfeld wrote: > Hello Viktor, > > Am 27.12.2015 um 19:41 schrieb Viktor Trojanovic: >> When I provisioned a new AD domain, I used the wrong network interface >> in the provisioning command. >> >> samba-tool domain provision --option="interfaces=lo eth2" --option="bind >> interfaces only=yes" --use-rfc2307

hi Michael Adam: Thanks for you work on samba. Here I am looking for some advice and your help. I have been stuck in continuous availability of samba 4.3.9 for two weeks. Continuous availability in SMB3 is an attractive feature and I am strugling to enable it. smb.conf, ctdb.conf are attached. Cluster file system is cephfs and mount to /CephStorage client: Windows 8 Pro root at node0:~# samba

Upgraded to Samba 4.11.2 and I?ve now too started seeing the message: get_share_mode_lock: get_static_share_mode_data failed: NT_STATUS_NO_MEMORY A lot. I modified the source in source3/locking/share_mode_lock.c a bit in order to print out the values of the service path, smb_fname & old_write_time when it fails and it seems they are all NULL? [2019/11/14 14:24:23.358441, 0]

Hello All, I am at the switch from sssd to winbind based samba domain members (Debian 9 stretch). I am using Samba 4.10.2 packages from Louis ( http://apt.van-belle.nl/ ) and rid backend for idmap. *My problem:* I am able to logon to my domain members using winbind_pam as long as my client is connected to a network where a domain controller is reachable. As soon as I shutdown and connect a