sybase I/O錯(cuò)誤，如何定位和解決問(wèn)題，謝謝

YeLLoW

在sybase數(shù)據(jù)庫(kù)上執(zhí)行sql語(yǔ)句，然后就hung住了，在log里面看到如下的問(wèn)題。個(gè)人覺(jué)得是有block損壞的情況，但是不能確定影響在哪個(gè)表上。
哪位同學(xué)可以幫忙定位一下問(wèn)題，并且提供一些解決問(wèn)題的方法，多謝。
00:00000:00000:2014/01/06 17:06:51.32 kernel  sddone: read error on virtual disk 7 block 2794496:
00:00000:00000:2014/01/06 17:06:51.32 kernel  sddone: 4096 bytes passed, 0 returned on read for virt disk 7 block 2794496
00:00000:00007:2014/01/06 17:06:51.33 server  Error: 1251, Severity: 26, State: 1
00:00000:00007:2014/01/06 17:06:51.33 server  An in-use preallocated semaphore cursor was encountered.
00:00000:00007:2014/01/06 17:06:51.33 server  Error: 1251, Severity: 26, State: 1
00:00000:00007:2014/01/06 17:06:51.33 server  An in-use preallocated semaphore cursor was encountered.
00:00000:00007:2014/01/06 17:06:51.33 server  Error: 1251, Severity: 26, State: 1
00:00000:00007:2014/01/06 17:06:51.33 server  An in-use preallocated semaphore cursor was encountered.
00:00000:00000:2014/01/06 17:07:51.32 kernel  sddone: read error on virtual disk 7 block 2794496:
00:00000:00000:2014/01/06 17:07:51.32 kernel  sddone: 4096 bytes passed, 0 returned on read for virt disk 7 block 2794496
00:00000:00007:2014/01/06 17:07:51.34 server  Error: 823, Severity: 24, State: 1
00:00000:00007:2014/01/06 17:07:51.34 server  I/O error detected during read for BUF=0xfffffd7df6575a18, MASS=0xfffffd7df6575a
18, Buf#=0, page=0xfffffd7bb2f6f000, dbid=4, Mass vdevno=7, vpage=2794496, Buf lpage=50037248, Mass stat=0x4000100, Buf stat=0
x1, size=4096, cid=0 ('default data cache'), Pinned xdes=0x0000000000000000, spid=0.
00:00000:00007:2014/01/06 17:07:51.34 server  Error: 1251, Severity: 26, State: 1
00:00000:00007:2014/01/06 17:07:51.34 server  An in-use preallocated semaphore cursor was encountered.
00:00000:00007:2014/01/06 17:07:51.34 server  Error: 1251, Severity: 26, State: 1
00:00000:00007:2014/01/06 17:07:51.34 server  An in-use preallocated semaphore cursor was encountered.
00:00000:00000:2014/01/06 17:08:51.32 kernel  sddone: read error on virtual disk 7 block 2794496:
00:00000:00000:2014/01/06 17:08:51.32 kernel  sddone: 4096 bytes passed, 0 returned on read for virt disk 7 block 2794496
00:00000:00007:2014/01/06 17:08:51.32 server  Error: 823, Severity: 24, State: 1
00:00000:00007:2014/01/06 17:08:51.32 server  I/O error detected during read for BUF=0xfffffd7df65755e0, MASS=0xfffffd7df65755
e0, Buf#=0, page=0xfffffd7bb2f6a000, dbid=4, Mass vdevno=7, vpage=2794496, Buf lpage=50037248, Mass stat=0x4000100, Buf stat=0
x1, size=4096, cid=0 ('default data cache'), Pinned xdes=0x0000000000000000, spid=0.
00:00000:00007:2014/01/06 17:08:51.32 server  Error: 1251, Severity: 26, State: 1
00:00000:00007:2014/01/06 17:08:51.32 server  An in-use preallocated semaphore cursor was encountered.
00:00000:00007:2014/01/06 17:08:51.32 server  Error: 1251, Severity: 26, State: 1
00:00000:00007:2014/01/06 17:08:51.32 server  An in-use preallocated semaphore cursor was encountered.
00:00000:00000:2014/01/06 17:09:51.31 kernel  sddone: read error on virtual disk 7 block 2794496:
00:00000:00000:2014/01/06 17:09:51.31 kernel  sddone: 4096 bytes passed, 0 returned on read for virt disk 7 block 2794496
00:00000:00007:2014/01/06 17:09:51.31 server  Error: 823, Severity: 24, State: 1
00:00000:00007:2014/01/06 17:09:51.31 server  I/O error detected during read for BUF=0xfffffd7df6575508, MASS=0xfffffd7df65755
08, Buf#=0, page=0xfffffd7bb2f69000, dbid=4, Mass vdevno=7, vpage=2794496, Buf lpage=50037248, Mass stat=0x4000100, Buf stat=0
x1, size=4096, cid=0 ('default data cache'), Pinned xdes=0x0000000000000000, spid=0.

Anxiang1129

這個(gè)是Sybase官方的錯(cuò)誤代碼及解決辦法。
http://infocenter.sybase.com/arc ... AdvRes/BACJACIC.htm

hobbylu

select name from master..sysdevices where low/16777216=7
然后檢查一下操作系統(tǒng)的日志以及該硬件的狀況

ceozenliu

哪種OS？還有ASE的版本？

YeLLoW

回復(fù) 4# ceozenliu
OS 是 solaris 10， ASE 是 15.0.3


   

YeLLoW

@Anxiang1129 & @hobbylu ，多謝，我看一下。

ceozenliu

(1) 你後端的storage應(yīng)該是array，先檢查disk array，看有沒(méi)有錯(cuò)誤燈號(hào)，有的話直接換掉那個(gè)硬盤(pán)，這是最幸運(yùn)的。

(2) 如果disk array沒(méi)有錯(cuò)誤燈號(hào)：
1>select phyname from master..sysdevices where vdevno=7   
2>go  
--可以看到是哪個(gè)物理儲(chǔ)存位置，換掉那一顆disk （這是指有RAID的情況）

一般而言，第(2)個(gè)方法應(yīng)該就可以解決問(wèn)題，如果還是不行
1>dbcc traceon(3604)
2>go
1>dbcc checkdb(db_name(4))   --如果數(shù)據(jù)庫(kù)大的話，會(huì)很耗時(shí)
2>go
--從dbcc checkdb的回傳訊息中，找看看有沒(méi)有「error」關(guān)鍵字，如果有，應(yīng)該就可以看到是哪個(gè)object有錯(cuò)，接下來(lái)就一個(gè)一個(gè)的修復(fù)

andkylee

設(shè)備vdevno=7有磁盤(pán)故障了。
檢查壞頁(yè)所屬對(duì)象：
dbcc traceon(3604)
go
dbcc page(4,50037248,1)
go
找到 object id ，然后select object_name(object_id)。

YeLLoW

謝謝樓上的各位同學(xué)的幫助，感覺(jué)收獲很多。
通過(guò)sysdevice表，發(fā)現(xiàn)出問(wèn)題device是用于記錄log的device。
由于項(xiàng)目上時(shí)間比較緊張，沒(méi)有時(shí)間進(jìn)行硬件排查，最后的解決方案就是drop 相關(guān)的db， 重建device，重新創(chuàng)建db。
目前問(wèn)題已經(jīng)解決，再次感謝大家。