On Tue, Sep 03, 2002 at 12:50:38AM +0900, Junjiro Okajima wrote: > > Claudio Jeker: > > Sorry that I missed your mail Gene. An anoying bug in both pbsadd and > > pbscheck mad it impossible to use multiple servers. I have debuged the > > problem and a fix will be comming soon. > > i am not sure what the fix "to use multiple servers" means, but i have > tried reading and understanding new pbs-tools in 20020901 patch. > there still be a bug in pbscheck, isnt it? after first sendrequest() > and timeoutread() call, the array named buf is filled with the answer > from the first pbsdbd, i.e. buf[0] is 'R' or something. before we try > next pbsdb, we must repair the buf, as putting 'Q' to buf[0]. >
No. After the first sendrequest and timeoutread the buffer is the same if there was no reponse. pbscheck tries only the other servers if the previous was down (timeout after one second). We assume that all pbsdbd servers have the same info stored (pbsadd sends the update/add messages to all pbsdbd servers). So if pbscheck recievs an answer there is no need to send a request to a other pbsdbd server. -- :wq Claudio
