We have a 2 node replication enabled setup (master/client). Frequently I'm seeing that DB fails to open on the client with the following error logs:
[1387876299:345435][693/139824764127264] REP_UNDEF: Read in gen 8
[1387876299:345707][693/139824764127264] REP_UNDEF: Read in egen 6
[1387876299:345809][693/139824764127264] REP_UNDEF: EID 0 is assigned for site 220.127.116.11:4998
Finding last valid log LSN: file: 1709 offset 4018264
What would cause this frequent error?
Can you provide some information about your application ?
At least, we'd like to know the BerkeleyDB version you are using and the platform your application is running on. Since different platforms and different versions can have different error information.
Winter, Oracle Berkeley DB
We are using DB version 5.1.19 on RHEL 6.3, with these patches:
# Intends to solve locked sync
patch -Np0 <../21036_5.1.19_1.diff
# Intends to solve PANIC: log sequence error by replication aware logging
patch -p0 <../20342_5.1.19_1.diff
# Intends to solve haging passive
patch -Np0 <../21326_5.1.19_msghang2.diff
# Intends to solve PANIC: unpinned page
patch -Np1 <../19130-5.1.19.diff
# Intends to fix patch 20342_5.1.19_1.diff: log files not removed
patch -Np0 < ../21448_5.1.19.diff
# Intends to add repmg queue limit and statistics.
# In this patch, we modified the IN_QUEUE_LIMIT_BULK from 1000 to 100000
patch -p0 < ../21420_5.1.19sv_3.diff
# Intends to solve PANIC: file size not a multiple of the page size
patch -p0 < ../21254_5.1.19.diff
# Intends to improve performance by more accurate locks
patch -p0 < ../21454_5.1.19_prelim3.diff
# The patch to fix DB_NOTFOUND panic during initial synch process
# NOTE. This patch is incompatible with the diagnosic 21314_5.1.19_repdiag3.diff patch.
# the diagnostic patch 22451_5.1.19_repdiag4.diff must be used instead of 21314_5.1.19_repdiag3.diff
# if 20668_5.1.19_1.diff is applied.
patch -p0 <../20668_5.1.19_1.diff