ORA-600(4400)错误

Oracle的600错误，错误号4400，这个错误和分布事务有关。

从alert文件中可以看到错误信息为：

Errors in file /opt/oracle/admin/data01/udump/data01_ora_22281.trc:
%uD.n*t
x
Wz0ORA-00600: internal error code, arguments: [4400], [48], [], [], [], [], [], []

不过从相应的trace文件中看不到导致问题的SQL语句：

ksedmp: internal or fatal error
"U,[2t*bZUR{0ORA-00600: internal error code, arguments: [4400], [48], [], [], [], [], [], []ITPUB个人空间t(EC
C2c.f6br
----- Call Stack Trace -----ITPUB个人空间i"v5K1x1e4KI%Q
calling              call     entry                argument values in hex     
xk0Mr-Y5r:t4]/Q0location             type     point                (? means dubious value)    ITPUB个人空间g)hQ9~*X2G)~&eb
-------------------- -------- -------------------- ----------------------------
(`B*Op;bl0ksedmp()+328         CALL     ksedst()+0           FFFFFFFF7FFFAB10 ?
2g2K;c%o L I6X7g.~HC0                                                   000000000 ? 000000000 ?
|Fa&{5iO-w|g$M0                                                   00000003E ?ITPUB个人空间B4J7B4VPH6a
                                                   FFFFFFFF7FFFB3A8 ?
y9rsM;R0                                                  1031C9458 ?ITPUB个人空间%O/R2c zB4@
kgeriv()+208         PTR_CALL 0000000000000000     000000000 ? 000103400 ?
;tb*T o&Y-H&ra0                                                   0001035CD ?000102C00 ?
,iJ!Q1Mcqv}0                                                   1035CD000 ? 1035CD328 ?ITPUB个人空间9Z%X fD.mA5f
kgeasi()+180         CALL     kgeriv()+0           1035CD588 ?1036F1108 ?ITPUB个人空间/?eQC;_+z
                                                   000000258 ?0000013C8 ?
&K z ^.jM.C}A0                                                   FFFFFFFF7FFFBC48 ?ITPUB个人空间9A(d)^(~Z?;_/?[&PE
                                                   1035CE958 ?ITPUB个人空间:RF1y V/W[5Q
Cds
ktcddt()+328         CALL     kgeasi()+0           1035CD588 ?1036F1108 ?
'@-H dyZ6{)\0                                                   000001130 ? 000000002 ?ITPUB个人空间 uA-f7J^)`
                                                   000000001 ? 000000004 ?ITPUB个人空间K#O0r;{ hI}[ e3J?
kssdch_stage()+484   PTR_CALL 0000000000000000     400DD283C40 ? 000000001 ?ITPUB个人空间
mme"i8h
                                                   000000001 ? 1035CD328 ?ITPUB个人空间o,EnW5v2],Z
                                                   000000030 ?00000000C?
1IV6i4_$?_/w$gZ0ktcbod()+840         CALL     kssdch()+0           400DD321538 ? 1035CFF50 ?
%kYwYF0                                                  380007C2C? 000000000 ?
U_V+_j3|9G"z5Q0                                                   400DD283C50 ? 400DD321558 ?ITPUB个人空间%Js|$B8?)],nuW
kssdch_stage()+484   PTR_CALL 0000000000000000     400DD321538 ? 400DB5C1F90 ?ITPUB个人空间o~xx/_XZs(XuZ
                                                   1035CFBF0 ? 000000000 ?ITPUB个人空间*n|I(n0g$^
                                                   000000000 ? 07FFFFC00 ?

根据metalink的查询结果，发现这个错误和分布事务有关：Doc ID: Note:138822.1。

虽然看不到导致问题的SQL语句，不过继续查询trace文件，发现问题确实和分布式事务有关：

FFFFFFFF7FFFC0F0 65642074 72616E73 006374696F6E2069  [ed trans.ction i]
4u8y-W3nQ.j0FFFFFFFF7FFFC100 00000001 035CD6C8 73736174 74656D70  [.....\..ssattemp]
-d2Xl1Y5a o0FFFFFFFF7FFFC1107420746F20626567 696E0064 69737472  [t to begin.distr]ITPUB个人空间.qW[v'n bs%U
FFFFFFFF7FFFC120 69627574 65642074 72616E73 61637469  [ibuted transacti]
R u*l3Q a;q6KH&i0.
,T9QA6QV[_0.
]e.Mwd8X0B0.
[`Z'}7Mx"gp;f0FFFFFFFF7FFFC7E0 FFFFFFFF 000000004F52412D 30323034  [........ORA-0204]
t(a{B D7Nt0FFFFFFFF7FFFC7F0363A2064 69737472 69627574 65642074  [6: distributed t]
q-j Dr:@:e&i+dD{0FFFFFFFF7FFFC800 72616E73 616374696F6E20616C726561  [ransaction alrea]ITPUB个人空间S-j:WX'F.F5vHe
FFFFFFFF7FFFC810 64792062 6567756E0074696F6E000000  [dy begun.tion...]

进一步查询METALINK，发现了下面的文章对这个问题进行了比较详细的说明：Doc ID: Note:464861.1。

根据文档的描述，跨版本的数据库链访问就可能碰到这个问题，Oracle在10.1.0.3中已经解决了这个bug。

对于这个问题，可以简单的忽略掉，不会造成任何的实际危害。