Thank you Aurelien, I should have RTFM more closely. Also I've ran strace -c which is claimed to report time spent in kernel...
on that "slow" system it is % time seconds usecs/call calls errors syscall ------ ----------- ----------- --------- --------- ---------------- 43.25 3.322673 10008 332 171 stat 25.46 1.955775 7190 272 28 open 16.58 1.273361 6006 212 162 access 12.25 0.940696 649 1449 1 lstat 1.06 0.081252 16250 5 1 unlink 0.78 0.060004 30002 2 link whenever on another node (with no problem): % time seconds usecs/call calls errors syscall ------ ----------- ----------- --------- --------- ---------------- 66.35 0.008808 0 18427 read 9.08 0.001206 0 52497 rt_sigprocmask 5.73 0.000761 22 34 clone 3.53 0.000468 1 342 180 stat 3.48 0.000462 0 2354 20 close 3.29 0.000437 0 908 28 open 2.83 0.000376 0 1410 1 lstat 2.45 0.000325 5 68 34 wait4 > reassign 517122 linux-2.6 > retitle 517122 linux-2.6: very slow access/open/... syscalls on NFS mounted > files > thanks > > whenever on other nodes it takes just 0.0001 or so > This time is the time the userland (ie libc6) waits before getting an > answer. -- Yaroslav Halchenko Research Assistant, Psychology Department, Rutgers-Newark Student Ph.D. @ CS Dept. NJIT Office: (973) 353-1412 | FWD: 82823 | Fax: (973) 353-1171 101 Warren Str, Smith Hall, Rm 4-105, Newark NJ 07102 WWW: http://www.linkedin.com/in/yarik -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org