[KLUG Members] Runing Out of Re:'s on Samba reprocess config
file question
Adam Williams
members@kalamazoolinux.org
Mon, 11 Mar 2002 10:14:10 -0500 (EST)
>>>Until I can figure out how to migrate my users to the new Monarch server
>>>running on RedHat 7.2, my Samba server is Samba 1.9.18p5 ala May 1998
>>>running on Red Hat 5.1 Kernel 2.0.36.
>>Yikes!!! Samba 1.9.x doesn't "support" anything much newer than NT4.0sp3.
>>Running an old Samba with new/updated clients is as perilous as meddling
>>in the
>>affairs of wizards. M$ is constantly "tweaking" and extending the CIFS
>>protocol and their RPC suite, newer Samba's are aware of the tweaks and
>>adjust accordingly.
>Fortunately, most of my users are stuck on Win98 SE and Lotus v5 with no
>immediate path out.
>I am supposed to be leading the migration from Lotus 123 to integration into
>our M$ based accounting package that has manufacturing modules. It is
>running on NT 4.0 sp6+ with a betrieve database engine, but for our
Ack!
>production app it looks like we have to move everything to M$ SQL server and
>re-write the entire app so it works with the accounting and manufacturing
>modules (Oh, Joy!!! ).
M$-SQL-2000 is not a terrible database, not great, but probably one of
their better products.
There are *LOTS* of UNIX and client/server apps with manufacturing
modules. If you're looking at a re-write I'd examine options. I got a
message (quasi-SPAM) from a firm near GR that has announced support for
Linux.
>>I've seen M$ clients lock read-only files (even executables!). Use
>>smbstatus to see if any funny locking is going on.
>Unfortunatley, I don't know what funny locking is. Smbstatus was my first
>approach to figure out what was going on, but I have trouble understanding
>what the various combinations of DenyMode, R/W and Oplock mean for the
>various smb clients, so I pretty much gave that up.
Yes, it is confusing. That is part of the problem. The clients confuse
themselves!
One has to understand that an oplock != a traditional lock. An oplock
lets me cache a chunk of a remote file locally. It is really a way of
registering for change notification. The server notifies the client,
"toss your cache" the file has changed beneath you. It is a little more
complicated than that, and one can see how that whole system
(non-cooperative clients, etc...) can get real dicey.
A lock restricts the access to a file (or region of a file) permitted by
other processes.
Then there are "second level" oplocks....
>For example, I have seen EXCLUSIVE+BATCH Oplocks on DENY_NONE RDONLY files
>while another file will have Oplock NONE on DENY_ALL RDWR. This makes no
>sense to me.
I have a good URL on all this lying about somewhere, I'll see if I can
dig it up.
>s it up to Samba to decide the type of Oplock or does the app ask Samba for
>a specific type of Oplock? I have read most of the man pages, docs and
>online books I can find about Samba, but I just don't get it.
>>Do you have any "process no longer exists" messages in your log file? We
>>have
>>Lotus users, and Lotus exhibits the most bizarre file management
>>techniques.
>>We've actually unconvered and reported Samba bugs based upon Lotus 123
>>Millenium. (It ain't much better sharing from a true M$ file server).
>No, I didn't see any process no loger exists, messages, but I have seen some
>really bizare entries where an SMBopenX fails and the next transaction opens
>the file that doesn't exist (and then starts to write to it)!!!
>For example I logged this last summer when several people were having samba
>access problems.
> stat of jwhitcomb/control data/1c6a1010 failed (No such file or directory)
> 2001/07/19 11:09:31 error packet at line 783 cmd=8 (SMBgetatr) eclass=1
> ecode=2
> error string = No such file or directory
> 2001/07/19 11:09:31 Transaction 30715 of length 102
> switch message SMBopenX (pid 25928)
> chdir to /home/account
> chdir to /home/account
> chdir to /home/bmd
> unix_clean_name [./JWHITCOMB/CONTROL DATA/1C6A1010]
> unix_clean_name [jwhitcomb/control data/1c6a1010]
> Allocated new file_fd_struct 9, dev = ffffffff, inode = ffffffff
> 2001/07/19 11:09:31 kblum opened file jwhitcomb/control data/1c6a1010
> read=Yes write=Yes (numopen=10 fnum=76)
> set_share_mode: Created share record for jwhitcomb/control data/1c6a1010
> (dev 2065 inode 2674712)
> set_share_mode: Created share entry for jwhitcomb/control data/1c6a1010 with
> mode 0x12 pid=25928
Is the file in control data a worksheet or some temportary file? If the
file gets created what is the case (upper/lower) of the filename?
>>Set level 10 logging for one machine and record the session.
>I shall do so Monday.
>And I shall try to get that new Samba server up ASAP.