Hi,

my problem are diskless pcs with aufs-rootfs and nfs4-mounted homedirectories.
newest aufs.

1) If the homedir is on our netapp (onTap 7.2.2) an utimes call is delayed by 
exactly 1 second.
2) If the homedir is on a linux nfs4 server an utimes call is not delayed.
3) If mounting the netapp homedir from a standalone linux without aufs the 
utimes call is also not delayed.

Here some test output:

#########################################################################
$ grep $HOME /proc/mounts; strace -tte close,open,utimes touch a
+ grep /home/np9 /proc/mounts
netapp1.public.ads.uni-freiburg.de:/vol/home/np9 /home/np9 nfs4 
rw,nosuid,nodev,vers=4,rsize=65536,wsize=65536,hard,intr,proto=tcp,timeo=600,retrans=3,sec=krb5i,addr=netapp1.public.ads.uni-freiburg.de
 0 0
+ strace -tte close,open,utimes touch a
18:43:51.955285 open("/etc/ld.so.cache", O_RDONLY) = 3
18:43:51.956440 close(3)                = 0
18:43:51.956710 open("/lib/librt.so.1", O_RDONLY) = 3
18:43:51.957628 close(3)                = 0
18:43:51.957893 open("/lib/libc.so.6", O_RDONLY) = 3
18:43:51.958657 close(3)                = 0
18:43:51.958917 open("/lib/libpthread.so.0", O_RDONLY) = 3
18:43:51.959828 close(3)                = 0
18:43:51.961066 close(0)                = 0
18:43:51.961351 open("a", O_WRONLY|O_NONBLOCK|O_CREAT|O_NOCTTY|O_LARGEFILE, 
0666) = 0
18:43:51.962409 utimes("/proc/self/fd/0", NULL) = 0
18:43:52.964600 close(0)                = 0
18:43:52.965485 close(1)                = 0
18:43:52.965838 close(2)                = 0
Process 6871 detached
#########################################################################
# tcpdump -n host 10.4.6.1
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on eth0, link-type EN10MB (Ethernet), capture size 96 bytes
18:43:51.961606 IP 132.230.1.204.469723390 > 10.4.6.1.2049: 304 getattr [|nfs]
18:43:51.962110 IP 10.4.6.1.2049 > 132.230.1.204.469723390: reply ok 660 
getattr [|nfs]
18:43:51.962120 IP 132.230.1.204.731 > 10.4.6.1.2049: . ack 1758740842 win 501 
<nop,nop,timestamp 707671 10283501>
18:43:51.962760 IP 132.230.1.204.486500606 > 10.4.6.1.2049: 272 getattr [|nfs]
18:43:51.963109 IP 10.4.6.1.2049 > 132.230.1.204.486500606: reply ok 160 
getattr [|nfs]
18:43:51.963208 IP 10.4.6.1.673 > 132.230.1.204.43693: S 341204666:341204666(0) 
win 8192 <mss 1460,nop,nop,sackOK,nop,wscale 0,nop,nop,timestamp 10283502 0>
18:43:51.963229 IP 132.230.1.204.43693 > 10.4.6.1.673: S 
3311647008:3311647008(0) ack 341204667 win 5792 <mss 1460,sackOK,timestamp 
707672 10283502,nop,wscale 7>
18:43:51.963340 IP 10.4.6.1.673 > 132.230.1.204.43693: . ack 1 win 8760 
<nop,nop,timestamp 10283502 707672>
18:43:51.963451 IP 10.4.6.1.673 > 132.230.1.204.43693: P 1:149(148) ack 1 win 
8760 <nop,nop,timestamp 10283502 707672>
18:43:51.963462 IP 132.230.1.204.43693 > 10.4.6.1.673: . ack 149 win 54 
<nop,nop,timestamp 707672 10283502>
18:43:51.963545 IP 132.230.1.204.43693 > 10.4.6.1.673: P 1:49(48) ack 149 win 
54 <nop,nop,timestamp 707672 10283502>
18:43:51.963618 IP 132.230.1.204.503277822 > 10.4.6.1.2049: 248 getattr [|nfs]
18:43:51.963661 IP 10.4.6.1.673 > 132.230.1.204.43693: F 149:149(0) ack 49 win 
8760 <nop,nop,timestamp 10283502 707672>
18:43:51.963679 IP 132.230.1.204.43693 > 10.4.6.1.673: F 49:49(0) ack 150 win 
54 <nop,nop,timestamp 707672 10283502>
18:43:51.963763 IP 10.4.6.1.673 > 132.230.1.204.43693: . ack 50 win 8760 
<nop,nop,timestamp 10283502 707672>
18:43:51.963867 IP 10.4.6.1.2049 > 132.230.1.204.503277822: reply ok 324 
getattr [|nfs]
18:43:52.003890 IP 132.230.1.204.731 > 10.4.6.1.2049: . ack 485 win 501 
<nop,nop,timestamp 707682 10283502>
18:43:52.964025 IP 132.230.1.204.520055038 > 10.4.6.1.2049: 272 getattr [|nfs]
18:43:52.964388 IP 10.4.6.1.2049 > 132.230.1.204.520055038: reply ok 336 
getattr [|nfs]
18:43:52.964415 IP 132.230.1.204.731 > 10.4.6.1.2049: . ack 821 win 501 
<nop,nop,timestamp 707922 10283602>
18:43:52.964997 IP 132.230.1.204.536832254 > 10.4.6.1.2049: 228 getattr [|nfs]
18:43:52.965179 IP 10.4.6.1.2049 > 132.230.1.204.536832254: reply ok 316 
getattr [|nfs]
18:43:52.965323 IP 132.230.1.204.553609470 > 10.4.6.1.2049: 252 getattr [|nfs]
18:43:52.965538 IP 10.4.6.1.2049 > 132.230.1.204.553609470: reply ok 340 
getattr [|nfs]
18:43:53.003947 IP 132.230.1.204.731 > 10.4.6.1.2049: . ack 1477 win 501 
<nop,nop,timestamp 707932 10283602>

25 packets captured
55 packets received by filter
0 packets dropped by kernel
#########################################################################



Any idea how to prevent this delay?

Best regards,
Martin


-- 
Martin Walter
University of Freiburg i.Br. --- Germany --- Fon/Fax: +49 761 203-4651/-4643
Rechenzentrum der Universitaet, Hermann-Herder-Str.10, D-79104 Freiburg i.Br.

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/

Reply via email to