We are using a replication set for a mongodb-installation. One day we realized that the replication has broken one database completely and we were unable to get the replication back to a synchronous state. Here is the stacktrace from the log-file:

c_prod Assertion failure a.ext()->xprev.isNull() src/mongo/db/pdfile.cpp 880
0xaf8c41 0xabf4cd 0x863eb9 0x86547b 0x865d50 0x6ed1c6 0x6e8321 0x6ea104 0x6eab4c 0x830028 0x83376b 0x7b0b0d 0x7b20e2 0x56fe42 0xae6ed1 0x7f177dd7c8ba 0x7f177d13002d
/usr/bin/mongod(_ZN5mongo15printStackTraceERSo+0x21) [0xaf8c41]
/usr/bin/mongod(_ZN5mongo12verifyFailedEPKcS1_j+0xfd) [0xabf4cd]
/usr/bin/mongod(ZN5mongo11freeExtentsENS_7DiskLocES0+0x679) [0x863eb9]
/usr/bin/mongod(_ZN5mongo6dropNSERKSs+0x6ab) [0x86547b]
/usr/bin/mongod(_ZN5mongo14dropCollectionERKSsRSsRNS_14BSONObjBuilderE+0x1a0) [0x865d50]
/usr/bin/mongod(_ZN5mongo7CmdDrop3runERKSsRNS_7BSONObjEiRSsRNS_14BSONObjBuilderEb+0xc6) [0x6ed1c6]
/usr/bin/mongod(_ZN5mongo12_execCommandEPNS_7CommandERKSsRNS_7BSONObjEiRNS_14BSONObjBuilderEb+0x51) [0x6e8321]
/usr/bin/mongod(_ZN5mongo11execCommandEPNS_7CommandERNS_6ClientEiPKcRNS_7BSONObjERNS_14BSONObjBuilderEb+0x12c4) [0x6ea104]
/usr/bin/mongod(_ZN5mongo12_runCommandsEPKcRNS_7BSONObjERNS_11_BufBuilderINS_16TrivialAllocatorEEERNS_14BSONObjBuilderEbi+0x2ac) [0x6eab4c]
/usr/bin/mongod(_ZN5mongo11runCommandsEPKcRNS_7BSONObjERNS_5CurOpERNS_11_BufBuilderINS_16TrivialAllocatorEEERNS_14BSONObjBuilderEbi+0x38) [0x830028]
/usr/bin/mongod(ZN5mongo8runQueryERNS_7MessageERNS_12QueryMessageERNS_5CurOpES1+0xc0b) [0x83376b]
/usr/bin/mongod() [0x7b0b0d]
/usr/bin/mongod(_ZN5mongo16assembleResponseERNS_7MessageERNS_10DbResponseERKNS_11HostAndPortE+0x3a2) [0x7b20e2]
/usr/bin/mongod(_ZN5mongo16MyMessageHandler7processERNS_7MessageEPNS_21AbstractMessagingPortEPNS_9LastErrorE+0x82) [0x56fe42]
/usr/bin/mongod(_ZN5mongo3pms9threadRunEPNS_13MessagingPortE+0x411) [0xae6ed1]
/lib/libpthread.so.0(+0x68ba) [0x7f177dd7c8ba]
/lib/libc.so.6(clone+0x6d) [0x7f177d13002d]
Wed Dec 12 16:35:08 got signal 15 (Terminated), will terminate after current cmd ends
Wed Dec 12 16:35:08 [interruptThread] now exiting

It seems like that there is one file / index that is causing that error and will force the synchronization to break.