I noticed that the OMC suspension was saturating but couldn't immediately see why (no ASC output or integrators left on, sliders haven't moved recently). I trended back to see when the suspension moved and found that it coincided with the Guardian machine reboot (alog90824), and trending more things I saw that the inputs to the lock filters went quickly from some thousands of counts to zero. These come from the M2 stage "ISC input filters," which also at the time of the Guardian reboot went from some thousands of counts to zero. Sheila suggested just moving the OMC back to its OSEM positions before the reboot, so that's exactly what I did. The OMC suspension is no longer saturating.