Hi Andy,
On Mon, Aug 23, 2004 at 06:07:14PM -0700, Andrew Hanushevsky wrote:
> I found the reason for the non-communications. When the olbd gets itself in
> the state where the server it has just selected has just gone down and no
> alternatives are available, it leaves a critical lock locked. This causes
> the system to essentially stop outbound communications. Most inbound
> communications can still be handled. All of the servers notice this and try
> to re-establish communications, which they do only to find that the olbd
> refuses to speak to them. Anyway, I will have a fix and that should generate
> a new release. We can restart the redirectors, at least.
Ok, I'll make a new release in a bit. Do you understand why this particular
behaviour wasn't seen in the past? Was it simply because you restarted things
on the redirector and then one-by-one were restarting things on the data
server?
Pete
-------------------------------------------------------------------------
Peter Elmer E-mail: [log in to unmask] Phone: +41 (22) 767-4644
Address: CERN Division PPE, Bat. 32 2C-14, CH-1211 Geneva 23, Switzerland
-------------------------------------------------------------------------
