Yes that is what I am seeing. I was wondering if there was a way to tweek the timeout.
-----Original Message----- From: Jeffrey Altman [mailto:[EMAIL PROTECTED] Sent: Friday, December 14, 2007 11:07 AM To: Jerry Normandin Cc: [email protected] Subject: Re: [OpenAFS] Working with 3 vlservers, I purposely disabled one to test queries take longer with afs01 removed Jerry Normandin wrote: > Hi all, > > > > I didn't see anything written on a FAQ so I've got a question for > you. Now that AFS is quick I've been I purposely disabled my primary > vlserver. > > SOme queries to the VLDB are fast, but 20% of them take a long time. Are > there any values I can tweek to reduce the wait for the other vlservers > to answer? The clients take the list of vlservers, sort them by their distance in the IP address space and then randomize them a bit. The end result is that if you have three VLDB servers, one third of your clients should have each one first in the list. For the third in which the down server is first in the list, the clients will have to timeout the server before it will fail over to the second server in the list. Is that what you are seeing? or something else? _______________________________________________ OpenAFS-info mailing list [email protected] https://lists.openafs.org/mailman/listinfo/openafs-info
