Mystery solved, I suppose.

Post #427,472 by mmoffitt 2/11/19 9:41:37 AM 2/11/19 9:41:37 AM Reply	Mystery solved, I suppose. Our SA said he spent the weekend on the postgresql mailing list. He said that after a lot of discussion, other folks had observed this when their database servers ran out of disk and that what I'd done (pg_dump and restore) was the only way to make things nice again. That was another detail that got erased from my op. This was brought to my attention only after the server had run out of disk. Thanks all. bcnu, Mikem It's mourning in America again.
Post #427,492 by crazy 2/13/19 11:58:13 AM 2/13/19 12:00:23 PM Reply	As I was reading I thought: exp/imp In my oracle days the only thing that would fix this is a "reorg", ie: exp to file, destroy/recreate, and imp back. Is this a highly updated database? Do records get bigger? Is there a lot of fragmented records and chaining? Edited by crazy Feb. 13, 2019, 12:00:23 PM EST Expand All History
Post #427,515 by mmoffitt 2/14/19 11:03:16 AM 2/14/19 11:03:16 AM Reply	I know very little about this db server. It's managed by another group who only calls me when they can't figure something out. There is a lot of updating that goes on I'm told, but I have no metrics. I'd never so much as logged into that host until this problem came up. There are a lot of little fiefdoms around here. ;0( bcnu, Mikem It's mourning in America again.

Welcome to IWETHEY!