Why does MaintainSwapSpace run continuously when swap is full?

I am running Web Polygraph against Squid (although I don't think this is a
Poly-specific problem), and I notice that as soon as the swap fills up, the
MaintainSwapSpace task runs, but since nothing in the cache is expired it
doesn't remove any objects. Since the cache is still full, it reschedules
itself to run immediately, consuming all available CPU time. Squid
continues to handle requests during this time, but I don't understand why
MaintainSwapSpace has this behavior. Is this a bug or a misconfiguration?

This happens with both Squid 2.3-STABLE3 and 2.4-DEVEL2 on Red Hat 6.2.

Here's an excerpt from the cache.log while this is happening:

2000/05/26 09:47:44| eventRun: RUN ID 1269789
2000/05/26 09:47:44| eventRun: Running 'MaintainSwapSpace', id 1269788
2000/05/26 09:47:44| eventAdd: Adding 'MaintainSwapSpace', in 0.000000
seconds
2000/05/26 09:47:44| storeMaintainSwapSpace: f=1.000000, max_scan=500,
max_remove=80
2000/05/26 09:47:44| storeMaintainSwapSpace: scanned 500/500 removed 0/80
locked 0 f=1.000
2000/05/26 09:47:44| storeMaintainSwapSpace stats:
2000/05/26 09:47:44| 156786 objects
2000/05/26 09:47:44| 500 were scanned
2000/05/26 09:47:44| 0 were locked
2000/05/26 09:47:44| 0 were expired
2000/05/26 09:47:44| comm_poll: 0 FDs ready
2000/05/26 09:47:44| comm_poll: time out: 959352464.
2000/05/26 09:47:44| eventRun: RUN ID 1269790
2000/05/26 09:47:44| eventRun: Running 'MaintainSwapSpace', id 1269789
2000/05/26 09:47:44| eventAdd: Adding 'MaintainSwapSpace', in 0.000000
seconds
2000/05/26 09:47:44| storeMaintainSwapSpace: f=1.000000, max_scan=500,
max_remove=80
2000/05/26 09:47:44| storeMaintainSwapSpace: scanned 500/500 removed 0/80
locked 0 f=1.000
2000/05/26 09:47:44| storeMaintainSwapSpace stats:
2000/05/26 09:47:44| 156786 objects
2000/05/26 09:47:44| 500 were scanned
2000/05/26 09:47:44| 0 were locked
2000/05/26 09:47:44| 0 were expired
2000/05/26 09:47:44| comm_poll: 0 FDs ready
2000/05/26 09:47:44| comm_poll: time out: 959352464.
2000/05/26 09:47:44| eventRun: RUN ID 1269791
2000/05/26 09:47:44| eventRun: Running 'MaintainSwapSpace', id 1269790
2000/05/26 09:47:44| eventAdd: Adding 'MaintainSwapSpace', in 0.000000
seconds
2000/05/26 09:47:44| storeMaintainSwapSpace: f=1.000000, max_scan=500,
max_remove=80
2000/05/26 09:47:44| storeMaintainSwapSpace: scanned 500/500 removed 0/80
locked 0 f=
2000/05/26 09:47:48| WARNING: Disk space over limit: 2188830 KB > 1048576
KB
2000/05/26 09:47:58| WARNING: Disk space over limit: 2188830 KB > 1048576
KB
2000/05/26 09:48:08| WARNING: Disk space over limit: 2188830 KB > 1048576
KB

Wesley Felter - wesf@us.ibm.com
Received on Fri May 26 2000 - 09:20:31 MDT