On Mon, 2009-11-16 at 08:46 +0100, Martin Nagy wrote: > Simo Sorce wrote: > > While working on a patch to use failover in the ldap driver I found a > > few bugs and a few things I felt missing. > > > > Attached a patch to fix bugs and add a function to get back a server > > name from a fo_server structure. > > > > Simo. > > > > P.S: if you are interested in the failover patch I have a working but > > unfinished patch here: > > http://fedorapeople.org/gitweb?p=simo/public_git/sssd.git;a=commit;h=29830482f5e44b5425fb91f82fd5a4ee692f3ae2 > > Just one thing. I was counting on the fact that the user of the fail > over will make sure that he doesn't try to create the same service > twice and doing so will mean a bug in the code. Looking at your patch > and having done a very quick look at your branch, it seems like you > count for this behavior (which I think is fairly reasonable). In that > case, could you also please change this debugging statement? > > DEBUG(1, ("Service %s already exists\n", name)); > > Please either completely remove it, or (IMHO better) change the debug > level.
I will probably raise the debug level, but it's not a big deal for now. > Regarding the ares error code, I'm not 100% sure what's the right > thing to do. Perhaps we should use a constant, like EIO and count on > the logging statement few lines up to be sufficient for the user. Any > other ideas? Not sure yet, what we need from callers are 3 type of errors. 1. an error that tells us that no server is resolvable/available (to avoid looping forever) 2. an error that tells us the current server is not resolvable/available 3. any other error should mean something really bad happened and we should abort. We miss n.2 for now so I am treating any error except ENOENT to mean the current server is not resolvable and we need to try the next one. Simo. -- Simo Sorce * Red Hat, Inc * New York _______________________________________________ sssd-devel mailing list sssd-devel@lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/sssd-devel