Disk Cache

Emergency Procedure

If there is a firm-wide or other emergency situation where cache integrity or loss of modeling results is a concern:

• Call or email us right away
• Send stack and heap dumps if appropriate (ie, you can skip this if there's a well-known cause like BBG disconnect)
• Attempt to recover:

   # Locate or build a cache that might work
   # Pick one user to try it out on
   # Restart Amoeba, choosing the replacement cache on startup
   # Do not change other settings!

If this fails:

• Tell us what happened

   # Send fresh stack and heap dumps, plus logs
   # Don't mess with the Amoeba instance until we follow up to get more info
   # Feel free to try steps 3-4 with a different user

• If it succeeds, notify us and repeat for everyone else
• Otherwise, we will advise the next step as soon as possible