Capture errors "10407 - data too large for queue buffer in record with rowid" during upgrade from (42398)

After upgrading to Shareplex 5.3 from Shareplex 5.0 or 5.1 you see the following errors in the event log:

.
04/28/08 10:06 Notice: detecting invalid numeric value 0x6c, default 0 is used [sp_ocap/8105]
04/28/08 10:06 Notice: data conversion failed [sp_ocap/8105]
04/28/08 10:06 Error: process sp_ocap killed due to SIGSEGV [sp_cop/23693]
.
.
.
04/28/08 10:08 Notice: invalid numeric length 254 for c2 46 08 ... [sp_ocap/10777]
04/28/08 10:08 Notice: data conversion failed [sp_ocap/10777]
04/28/08 10:08 Internal error: 10408 - data conversion error on rec with hrid:AAAfSGAAmAABMmgAAg [sp_ocap/10777]
.
.

04/28/08 10:08 Notice: data conversion failed [sp_ocap/10777]

04/28/08 10:08 Internal error: 10407 - data too large for queue buffer in record with rowid AAAfSGAAmAABMmgAAq [sp_ocap/10777]

04/28/08 10:08 Notice: invalid numeric length 56 for c2 46 08 ... [sp_ocap/10777]

Â

The latest *ocap log in $SP_SYS_VARDIR/log has the following entries:

08-04-28 10:06:04.077713 8105 1 Warning: detecting invalid numeric value in column data
08-04-28 10:06:04.077920 8105 1          tmp=-7,i=2,n=4
08-04-28 10:06:04.077979 8105 1          Use default value of zero to work around.
.
.
08-04-28 10:08:58.088090 10777 1 Data converison error, transop listing:
08-04-28 10:08:58.088245 10777 1 seqno/offset: 31757/18286668
.
.
08-04-28 23:51:01.526275 13098 1 data too large: size=83951642, space=304170, cpart=3
08-04-28 23:51:01.526410 13098 1 data too large: cpart=3, col=4, len=83951642, type=1
08-04-28 23:51:01.526480 13098 1 data too large: size=167772530, space=304126, cpart=5
08-04-28 23:51:01.526548 13098 1 data too large: cpart=5, col=6, len=167772530, type=1
.
.