PC name change

[Bok]

With some of the machines I brought home from work, I've manually changed their names..

All linux, this is causing some confusion on my fadprogress output as it reports two different names coming from each pc.

For instance one was originally called Blade9 at work, it is now Blade23 at home.

I use fadprogressd on this machine with a -name of Blade23.29 (29 is it's ip address). I get two lines from fadprogressd with Blade23.29 in the server piece and the names are

Blade9 & Blade23 respectively.

I can delete the Blade9.tlg file but it then gives a warning. Which other files should i edit in the fad directory to remove the reference to Blade9?

:cheers:

Bok

[PCZ]

Is this machine a fad server ?
I guess it is.

look for node.txt
It will have the old machine name in it.

Warning if you change this you will break the FAD security and the server won't start.
You will need to run fadsetup again and a new node id will be assigned.

[Bok]

It is a server, but I've already ran fadsetup again a few days ago... perhaps it's just looking through the server.log ? That still has entries for the old name ?

Bok

[MerePeer]

Hmm a bit hard to tell from your post but lets start with this, try:
./server -list

This is a dump of the server.dat file. You will see lots of jobs in the output and you'll see both the old node name and the new. In each line there are two dates, after the second date is a code: 02 means uploaded, 5 means in progress. It is most likely that you will see a 5 beside a record with the old node name. It is slightly possible you will also see a job listed as "Queued" for the old node name. What you need to do is remove the job that the server is listing as in progress (5) for that old node name (but not the new node name), which is done like this if you grab the job number from the left side of the output dump: (I cant recall if you keep the "j" in "j100")

./server -remove j100

It wouldnt hurt to remove the queued one although it is currently irrelevant to fadprogressd.

One alternative to -remove is to use ./server -reset j100 and I think that just makes it available for your new node to process.

You shouldnt need to restart fadprogressd after these chgs.

The only other file fadprogressd looks at, once every 10 mins, is the {name}.TLG file to get the cpu rating: but it will only look for & at that file if it finds that {name} in the server.dat with a status 5.

[Bok]

Ok,

that makes sense. When I try it though, it says another copy of server is already running. If I kill that, then run the remove, do I then just restart ./server ?

:cheers:

Bok

[MerePeer]

I think you are saying that the ./server -list worked but the -remove and other cmds dont unless you killall server. In that case, yes after you are done straightening out the queue aka server.dat, I usually just use ./loader and that seems to startup whatever isnt started and return control to me.
<edit> see PCZ response for additional details <end edit>

[PCZ]

You can leave think running while you mess with the queue.

It is a bit annoying that you have to kill server before any of the queue commands work though.

So stop server do all the queue manipulation then restart server.

./server &

to manually restart think
./think /n 0 &
./think /n 1 &
etc
increment 'n' for however many cpu's you have.