Yes, I certainly hope you have DB and Repository backups from just before you ran REPAIR.
Why did you run REPAIR if your projects were running ''fine''?
There is no need to do that unless HP-support asks you to -OR- when upgrading. When upgrading to a newer version of ALM, you do the REPAIR on a RESTORED project in the NEW version of your ALM instance.
This ensures it is repaired with the criteria of the new version, when complete (successful), you then UPGRADE the project. Usual process is: copy project DB to NEW DB instance, copy repository to new repository location, launch NEW ALM instance, then RESTORE / VERIFY / REPAIR/ UPGRADE the project.
So, for your dilemma... when a REPAIR fails, the only option is to RESTORE YOUR PROJECT DB from backup (on top of the corrupt one). For good measure, I usually also restore the entire project REPOSITORY as well from the same backup timeframe, just to be safe.
Moral of the story: DO NOT RUN REPAIR UNLESS YOU ARE UPGRADING A PROJECT and have backups, and do it on the NEW ALM server instance. (Of course, if Support asks you to on a case, it is OK to do otherwise.)