due to moving rooms around, my rigs are no longer permanently connected to the server.
when they are, WU’s are allocated to the rigs by name and transmission via CAT5 cabling takes a second or less.
now they are connected only every few days to upload results/download new WU’s, they take upto 7 mins per WU to download and are allocated to machines eg 123.456.789.000
even at 10Mbs/per sec rather than 100Mbs via CAT5 and 6 machines running data both ways, surely these transfer times are excessive?
or would it be limited by the network card in the server?
Can’t answer your question but SETI units are only 340-350Kbytes. or 2.8Mbits a unit so it should be faster than that
so something is throttling it i’d doubt it was the network cards.
Nothing strange about it Keith, it’s the way NetBIOS name browsing is handled by M$ OSes that’s causing the problem.
If you’re going to persist in plugging/unplugging machines from the network then you need to either handle all your addressing directly via IP address or give all the PCs static addresses and set up hosts/lmhosts tables on each one.
Edit the ‘hosts’ file (if its missing than rename the ‘hosts.sam’) and write all your IP’s with the computer names into it.
Put a return behind the last line (bug in W2k). Copy that file onto each of your computers. Restart them all.
If you are using lanmanager/netbios than do the same with the ‘lmhosts’ file.
After that resolving of Name/IP should work w/o probs.
Originally posted by JUGGY You shouldn’t need to reboot the machines as the hosts file updates dynamically (the OS check it for every request).
Basically right as long as you edit the files directly.
M$ is first checking for a change of the file and uses the table already in memory as long as there is no change of the file.
If you are copying the files, M$ sometimes doesn’t realize the change and as a result isn’t updating the IP/name table in memory (bug in W2k?)