Heritrix: Internet Archive Web Crawler

After only 20-30 second of running a crawl, URLs with
-2 error code and 10 tries (10t) show up in crawl.log.

Some crawl configuration used:
max-tries = 10;
retry-delay-seconds = 900
timeout-seconds = 1200
sotimeout-ms = 20000

First line of the crawl.log:

20040303013639842 1 55 #1 dns:www.cecc.gov
2132 text/dns
P http://www.cecc.gov/
.
.
.

About 20s later:

20040303013658845 -2 . #4
http://www.csce.gov/images/map-lft.gif . . 10t
E http://www.csce.gov/helsinki.cfm
20040303013658849 -2 . #3
http://www.csce.gov/images/text-search.gif . . 10t
E http://www.csce.gov/helsinki.cfm
20040303013658853 -2 . #2
http://www.csce.gov/images/menu-privacy.gif . . 10t
E http://www.csce.gov/helsinki.cfm
.
.
.
.

From local-errors.log (this images is retried 10 time
in just 3 seconds) :

First try:
20040303013649559 -2 . #4
http://www.csce.gov/images/map-lft.gif . .
E http://www.csce.gov/helsinki.cfm
java.net.SocketException: Socket is closed
at java.net.Socket.setSoTimeout(Socket.java:918)
at
org.apache.commons.httpclient.HttpConnection.setSoTimeout(HttpConnection.ja
va:623)
at
org.apache.commons.httpclient.MultiThreadedHttpConnectionManager$HttpConnec
tionAdapter.setSoTimeout(MultiThreadedHttpConnectionManager.java:1174)
at
org.apache.commons.httpclient.HttpClient.executeMethod(HttpClient.java:658)

at
org.apache.commons.httpclient.HttpClient.executeMethod(HttpClient.java:529)

at
org.archive.crawler.fetcher.FetchHTTP.innerProcess(FetchHTTP.java:178)
at
org.archive.crawler.framework.Processor.process(Processor.java(Compiled
Code))
at
org.archive.crawler.framework.ToeThread.processingLoop(ToeThread.java(Compi
led
Code))
at
org.archive.crawler.framework.ToeThread.run(ToeThread.java:100)

Last try (same error) at timestamp:
20040303013653835