On Wed, 28 Jun 2006, Matthew Soffen wrote: > David Lee wrote: > > [...] > > Matt: that "now"... Do you mean it used to work in the past but no longer > > does so? If so, do you have any "before and after" time-bracketing? (Or > > what CVS update precipitated it?) > > > > My own environment is Solaris. Has "STONITHDBasicSanityCheck" ever worked > > for you on Solaris? > > > > [...] > > I believe BSC worked in 1.99 version of Heartbeat (but my memory isn't > what it used to be) Also, I didn't have a build/test cluster for so > long, we were into the 2.0 sequence before I got my test system rebuilt > I could always fetch out a 1.99 version and verify that BSC works ( I > believe that it does ).
Going back to 1.99 probably won't gain us much. I was simply wondering whether we had a reasonably closely dated pairing of "known good" and "known bad" states. The time span of 1.99 to present seems far too wide to be of much value in any CVS post-mortem work. Also... My understanding is that the 2.x fundamentally requires authentication in the IPC layer, and that BSC exercises this aspect. If so, then BSC could never have fully passed (could it?) until the time that I coded up the streams support in "lib/clplumbing/ipcsocket.c" (revs 1.169-1.172) in January this year. (Aside, for our non-Solaris readers: Solaris sockets (at least up to and including S9) don't have an associated authentication mechanism, hence our providing the streams alternative (for those who might be curious, this is also available on some other OSes) which has inherent authentication.) -- : David Lee I.T. Service : : Senior Systems Programmer Computer Centre : : Durham University : : http://www.dur.ac.uk/t.d.lee/ South Road : : Durham DH1 3LE : : Phone: +44 191 334 2752 U.K. : _______________________________________________________ Linux-HA-Dev: [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev Home Page: http://linux-ha.org/
