Lustre discuss - May 2006 - question about ptlrpc_expire_one_request

Hello clusterfs,

I''m using lustre 1.3.2 with gcc 3.3.3 on SuSE 9.0 against vanilla
2.4.24

When I run IOZone in an endless loop I sometime get this error messages:

 >LustreError: 4370:0:(client.c:816:ptlrpc_expire_one_request()) @@@ 
timeout (sent at 1097225851) req@ceb2dc00 >x30114/t22042 
o4->media-ost1_UUID@NID_192.168.9.11_UUID:6 lens 288/248 ref 3 fl 
?phase?:R/4/0 rc 0/0
 >LustreError: 4370:0:(import.c:106:ptlrpc_set_import_discon()) 
OSC_media5_media-ost1_MNT_media5: connection >lost to 
media-ost1_UUID@NID_192.168.9.11_UUID
 >LustreError: 4349:0:(recover.c:113:ptlrpc_run_failed_import_upcall()) 
Invoked upcall /usr/local/lustre/recover.sh >FAILED_IMPORT 
media-ost1_UUID OSC_media5_media-ost1_MNT_media5 NID_192.168.9.11_UUID


My revover upcall script then re-integates the missing server, but my 
question is why it
loses sometimes the connection? Maybe it has something to do with the
"high network load"?


regards,

Martin