similar to: smb gives many collissions

Andrej Borsenkow wrote: > Current aside, the February2000 was last modified on February, 2nd. Last > message there is http://us1.samba.org/listproc/samba/February2000/0042.html > The same is true for samba-technical, samba-cvs (did not check others) > Or is it a problem of my proxy? I don't think so, I see the same thing. Samba web-folks, could you check this out, please?

samba@samba.anu.edu.au wrote: > > SAMBA Digest 1853 > > For information on unsubscribing see http://samba.anu.edu.au/listproc > Topics covered in this issue include: > > 1) Re: long winded printing LARGE files soloution > by Heiko Nardmann <h.nardmann@secunet.de> > 2) Samba replacing NFS > by Jonathan Peterson

William E. Dent will be away from Wednesday August 30, 2000 to Tuesday September 5, 2000. Mail is being forwarded to wedent@home.com. Original Message Text follows: ---------------------------------------------------------- Send samba mailing list submissions to samba@lists.samba.org To subscribe or unsubscribe via the World Wide Web, visit

Chris Hines wrote: > I would like to use samba as an NT domain controller for NT 4 & Windows 9X > and posibly windows 2000. I have configured samba 2.0.7 and samba NTG as > domain controllers and they seem to work. > > We wish our users to use a single password accross UNIX & windows which > are copied from a central NIS map managed by the University. Some time in >

I've been suggested to get some logs and give a brief description of the problem I'm encountering with Cadds5 / Samba 2.0.6 and offer them to you , via http or ftp , to try to find a solution I 've collected the logs , can I send them with a description of the problem Thanks Stefano Colombo ( scolombo@cdmtc.it ) System / Network Engineer CDM Tecnoconsulting SPA v. M.L.King 38/2

Dear Dave, Doesn't this only work if you're using samba as a domain controller and the client says "net use h: /home" I've not got it to redirect "//server1/username" to "//server2/username" when the NIS map says "user server2:/path/&" Though if it could be done I'd be VERY keen. Tim "David Collier-Brown"

Hello All, Thanks to David Collier-Brown who pointed me to the oplocks setting, what was needed was the following:- oplocks = off Regards, Leo -----Original Message----- From: David Collier-Brown [mailto:David.Collier-Brown@canada.sun.com] Sent: Thursday, 15 June 2000 22:54 To: Leo Kliger Subject: re: Strange Samba/NT compatibility issue You just got caught by cleint-side caching. If you

> Date: Thu, 27 Jul 2000 08:25:44 -0400 > From: David Collier-Brown <David.Collier-Brown@canada.sun.com> > To: Luc Lalonde <llalonde@gre.ulaval.ca>, samba@samba.org > Subject: Re: conditional logon scripts > Message-ID: <39802A48.CD84F4C6@canada.sun.com> > MIME-Version: 1.0 > Content-Type: text/plain; charset=us-ascii > Content-Transfer-Encoding:

Julien Boppert <boppert@hdz-ima.rwth-aachen.de> wrote: > we have an Enterprise 450 with Solaris 2.6 and mounted NFS-shares from > an Sparc 20 (Solaris 2.6 too). On the Enterprise we run Samba 2.0.5a as > PDC and export the mounted Shares with Samba to our WIN NT 4.0 > Servicepack 3 Clients. Erk! It's generally a bad idea to import data via NFS and re-export it via Samba:

A colleague just wrote and said he's had a problem with Samba using a Cascade ("SLS") PDC. This is odd, as Jeremy tested that rather recently! The symptom reported was, with security = domain and password server = <the SLS server>: > . When I try to validate > a login on the samba box against the SLS SAM, I receive the

Ian Leonard wrote: > >> If I use smbmount and copy a 180k file, it transfers in about 0 seconds > >> (including unmounting). > >> > >> If I use smbclient and 'put' the file, it takes 14 seconds. I have heard > >> of an MTU problem but can't see any configuration options. Ian then experimented with the buffer-size option of smbclient, and

John Wilkes wrote: > I can't get samba to work reliably. Win98 machines can map a network > drive via Samba, but writing large files to the Linux box fails. I get > error messages that all are some variation on this ... > [2000/02/20 22:38:31, 6] lib/util_sock.c:write_socket(560) > write_socket(6,39) wrote 39 > [2000/02/20 22:38:31, 10] lib/util_sock.c:read_socket_data(468)

Craig Weatherhead wrote: | I'm just curious if there is a way to have samba be able to restart | a transfer of a file from the point that it left off? It's doable "below" Samba using the Samba VFS, by passing the start-point as part of a filename and having the VFS (which you'd have to write!) return the rest of the file. Say you were transferring

Jeremy Allison wrote:> > Nope, I don't think so. Look at the kernel mailing lists for comments > comments about ext2fs performance with 10,000+ files - it sucks. > The developers consider this a broken application decision (ie. storing > that many files in a directory for an application, rightly > or wrongly). Try doing a regular ls of a directory with that many files, >

A moderately-frequently asked question is "why do automounted directories in shares disappear?" The old answers were: 1) they timed out and auto-un-mounted, and 2) you shouldn't re-export NFS-imported stuff anyway. A new answer (0) was just suggested by a Sun techie: add a "browse" option to the automounter map entry, to make them appear even when they're not

Andrew Wilson wrote: | For us the Solaris ACL is being honoured by Samba but can result in less | permissions granted on destination files and directories than the ACL | stipulates, eg. | ACL on a parent directory "test" is as follows:- | mask:rwx | However a file or a sub-directory created beneath "test" will be granted | less permissions for user tuser than desired, eg.

Jonathan wrote: | have upgraded SAMBA 2.04 to 2.06 and server versions 2.6 to 2.7 Solaris. | After reloading SAMBA I get an internal error in my logs and users an not | connect even though it shows up in the network. Below is the error in the | log ... That's a failure in --- BOOL attempt_netbios_session_request(struct cli_state *cli, char *srchost, char *desthost, struct in_addr *pdest_ip) {

Oh, you already replied... Also you're on Solaris, so it's easier: you can email me truss and sar files. So, if I understand you correctly, the upgrade to Solaris 7 and 2.0.5a cut performance down noticeably. This is unexpected, as a colleague who does big benchmarks always checks out Samba against the old 1.9.18 version to make sure there hasn't been a reduction, and 2.0.7, at

Thomas LeZotte wrote: > Does anyone know if you can share a directory on a unix system, witch > will allow users in the "write list" to write after logging in. But if > a user is not in the "write list" then they are automatically logged in > as guest, with read only permissions. The write list will do what you want, but the user must be logged in first, as

Has anyone seen file corruption with Excel 97 SR1, on Samba? Scotty (scott@dork.com) is running 2.0.7 on SCO 5.0.4 on a Compaq Proliant 800, compiled on SCO 5.0.5 box. He writes, on comp.protocol.smb: > Oh dear. If this is the case does that mean my clients can't use Excel > 97 SR1 on Samba or is there a workaround? At the moment some of my > clients users are saving files to their C