Re: invalid page header
От | Tom Lane |
---|---|
Тема | Re: invalid page header |
Дата | |
Msg-id | 9520.1186150080@sss.pgh.pa.us обсуждение исходный текст |
Ответ на | invalid page header (Markus Schiltknecht <markus@bluegap.ch>) |
Ответы |
Re: invalid page header
|
Список | pgsql-general |
Markus Schiltknecht <markus@bluegap.ch> writes: >> Block 58591 ******************************************************** >> <Header> ----- >> Block Offset: 0x1c9be000 Offsets: Lower 12858 (0x323a) >> Block: Size 28160 Version 73 Upper 14900 (0x3a34) >> LSN: logid 627535472 recoff 0x3a693b22 Special 9506 (0x2522) >> Items: 3209 Free Space: 2042 >> Length (including item array): 8192 >> >> Error: Invalid header information. Hm, looks suspiciously ASCII-like. If you examine the page as text, is it recognizable? We've seen cases where, for instance, a page of a system's mail spool had replaced a database page. It's hard to be sure whether that sort of thing is a hardware error or an operating system bug. > What's the best cure? Can I just wipe out the block with something like: > # dd if=/dev/zero of=base/296788/302602 seek=58591 bs=8192 count=1 Yeah, try that. I think the net effect will be that some wide (toasted) values will be truncated. regards, tom lane
В списке pgsql-general по дате отправления: