[KLUG Members] SAMBA... problem or solution ....

[Robert G. Brown]

>From a novice SAMBA operator....

Perhaps both!

I inherited a network that consists of (among other things) a
SAMBA server (very recent version, like 2.2.7) and a gaggle of
Win98SE and WinXP clients, hosted on machines from 433 mHz to 
2.4 gHz.

At startup, some of the users complain of failure to connect to
their drive shares. Its easy to go into "My Computer" or some 
similar utility and complete the shares at that point. I've also
logged off, then back on, and the shares are completed without a 
problem. Reboot, and the problem returns.

All this sets off some thinking... clearly, lots of things go on
at startup... like drivers getting loaded, devices being initial-
ized, and so forth. Perhaps, the thinking goes, there is a timing
problem of some sort... like networking on the client is not quite
ready for the initiation of SMB activity.

Problems are almost/completely non-existant on the slower clients,
almost constant on the fast ones. This is not a really pure test,
since the slower ones are '98SE, while the faster ones are XP, so
we can't completely factor out OS and specific drivers.

Even with that in mind, I can't help feeling that introducing some
delay on one side or the other would be a useful way to explore
this idea... if I could slow down or delay SMB initiation on the 
client (even by, say, a second or so) it would prove (or disprove)
my thinking about these sorts of timing problems. Introducing some
sort of delay or retry on the SAMBA server may also prove interesting.

This leads to a number of questions for all you SAMBA sever veterans:

1. Have you seen this problem before? If so, is there a straightaway
   solution?

2. Do such timing/delay/retry control exist? On clients? On servers?

3. Am I barking up the wrong tree? Someone got a better tree?


I look forward to answers..
                                                Thanks and Regards,
                                                 ---> RGB <---