On 9/26/05, Ian mu <[EMAIL PROTECTED]> wrote: > -- > [ Picked text/plain from multipart/alternative ] > It would be doubtful another dist would help (unless the current one is > setup wrong which is possible), but its hard to tell without seeing more > info. I.e can you paste the output from "top" and "free" and the server > settings. Do any other processes use larger amounts of cpu than you would > expect? Is it just one server using loads, or do all the servers use a lot > of cpu? > > On 9/26/05, franck.turmel-josek <[EMAIL PROTECTED]> wrote: > > > > -- > > [ Picked text/plain from multipart/alternative ] > > Thank you for this help. > > > > We have done a lot of test with various sys_ticrate, sv_maxrate, > > sv_minrate... the problem is the same : lags with 'stats' CPU closes to 100% > > and 'top' CPU less than 50%. > > > > We will look for virtual CPU % use (not all in the same time) and 'stats' > > CPU info. > > > > Our box with Fedora Core 3 is with HT. > > Then if the lags happen when the virtual CPU is full, what can we do ? > > Nothing HLDS is like that ... ? > > Changing our OS for another linux distro ? > > > > Thx > > > > Franck > >
'vmstat' can be handy for this sort of thing as well. It will show you the number of processes in the run queue, how many swaps in and out are occuring, blocks in and out to and from disk drives, as well as a break down of how much userland, kernel, and IO wait is being consumed by processes. Try running 'vmstat 1' and you'll see an update every second of what's happening on your machine: Use this in addition to 'top' and anything else. Its not a replacement for other tools, but a nice addition in diagnosing busy systems. Sorry if someone else has suggested this already. -- Alex _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
