Problems running under hypertable

[Patrick Scott]

I get the following errors under heavy load from hypertable on top of
kfs:

09-06-2011 17:32:08.614 ERROR - (KfsOps.cc:2020) Write prepare failed,
lease expired for 844
09-06-2011 17:32:08.614 ERROR - (ClientSM.cc:135) I12236I
10.218.15.199:46710 seq: 150147 status: -1001 buffers: 62976 write-
prepare: seq = 150147 chunkId = 844 chunkversion = 1 offset: 62786048
numBytes: 62976 msg: lease expired took: 5e-05 sec.
09-06-2011 17:32:08.614 ERROR - (KfsOps.cc:2209) failed: lease expired
write-sync: seq = 150149 chunkId = 844 chunkversion = 1 offset =
62786048 numBytes 62976 write-id info: 10.218.15.199 30000
2556049131738106703
09-06-2011 17:32:08.614 INFO - (ChunkManager.cc:2291) Setting the
status of writeid: 2556049131738106703 to -1001
09-06-2011 17:32:08.614 ERROR - (ClientSM.cc:135) I12236I
10.218.15.199:46710 seq: 150149 status: -1001 buffers: 4096 write-
sync: seq = 150149 chunkId = 844 chunkversion = 1 offset = 62786048
numBytes 62976 write-id info: 10.218.15.199 30000 2556049131738106703
msg: lease expired took: 2.3e-05 sec.
09-06-2011 17:32:08.614 ERROR - (KfsOps.cc:2020) Write prepare failed,
lease expired for 844
09-06-2011 17:32:08.614 ERROR - (ClientSM.cc:135) I12236I
10.218.15.199:46710 seq: 150148 status: -1001 buffers: 3072 write-
prepare: seq = 150148 chunkId = 844 chunkversion = 1 offset: 62849024
numBytes: 3072 msg: lease expired took: 2.7e-05 sec.
09-06-2011 17:32:08.614 ERROR - (ChunkManager.cc:2270) Clone write op
failed due to status: -1001
09-06-2011 17:32:08.614 ERROR - (KfsOps.cc:2167) failed: no such write
id write-sync: seq = 150150 chunkId = 844 chunkversion = 1 offset =
62849024 numBytes 3072 write-id info: 10.218.15.199 30000
2556049131738106703
09-06-2011 17:32:08.614 ERROR - (ClientSM.cc:135) I12236I
10.218.15.199:46710 seq: 150150 status: -22 buffers: 4096 write-sync:
seq = 150150 chunkId = 844 chunkversion = 1 offset = 62849024 numBytes
3072 write-id info: 10.218.15.199 30000 2556049131738106703 msg: no
such write id took: 1.9e-05 sec.

Are there some tuning parameters to avoid errors like these? What
might be causing these problems? My setup is a single machine with
both the metaserver and a chunkserver. Hypertable is reading and
writing to kfs a lot if that helps.

Patrick