You could turn on traceoptions under OSPF and set a flag for whatever you are interested in logging. In most networks, this shouldn't generate too much data, so I suspect it would be fine to do longer-term if filtered appropriately.
I used the traceoptions on both of my VRs ospf configuration (one log file per VR). Currently using the general (general is a combination of the normal and route flags), events, packets, and error flags. Although, the packets flag logs all ospf related packets including hello messages, which may be more than some people want in their log files. You could remove the packets flag and replace it with just the specific packet flags named lsa-update, lsa-request, and lsa-ack. If you do this, your logs will be much smaller as the hello messages happen quite often. As a result, you could probably reduce the log file size from 1m to 100k or less.
file ospf_trace size 1m files 10 world-readable; flag error; flag general; flag event; flag packets;