Capture stops due to error and you see the following in the event_log:
04/08/04 10:31 Error: 13016 - Invalid magic number while reading object cache. [sp_ocap(osp)/3971]
04/08/04 10:31 Internal error: 10703 - can't read object cache o.atrn 42 [sp_ocap/3971]
04/08/04 10:31 Process exited sp_ocap (for o.atrn) [pid = 3971] - exit(1)
Although capture can stop due to Internal error: 10703 - can't read object cache, the next message will tell you why, such as no file or directory. In this instance, it is due to Invalid magic number. Somehow the capture object cache has become corrupt. If this case the object cache with activation number 42 has become corrupt. This can be due to simple corruption or by backing out to a lower patch or shareplex install level. This could be a result of doing something on the Oracle level that would interfere with the internal table of Shareplex.
Move the current ocap object cache to .old . Then copy the sp_conf objcache and rename it to the capture object cache file name.
If the resolution does not resolve the issue, please call Shareplex Support.