[rescue] UFS quota woes

Brian Hechinger rescue at sunhelp.org
Sun Jun 17 22:09:42 CDT 2001

so it was taking a REALLY long time to log into my one machine.  that's odd,
it's not overloaded, and once your in it's fine.  i thought maybe it could be
a reverse DNS issue, but it isn't, so i dug a little deeper.  looks like the 
part where /etc/profile runs /usr/sbin/quota was hanging it up.  i commented out
the line with /usr/sbin/quota and now i can login at normal speed.

i don't use quotas (my own machines, no-one else uses them, etc) so commenting
that line out is no problem, but i'm curious as to why quota hangs on that box
but not the other sparc20.

well, truss isn't being real useful since it won't let me run it on a suid
binary as not-root and root doesn't seem to suffer from this quota hanging
problem.  so i ran quota in one window and quick did a ps and truss -p as
root in another, and i got this small bit of info:

# truss -p 14827
    *** SUID: ruid/euid/suid = 100 / 0 / 0  ***
poll(0x0004024C, 1, 15000)      (sleeping...)
poll(0x0004024C, 1, 15000)                      = 0
putmsg(4, 0xEFFFF6A4, 0xEFFFF7BC, 0)            = 0
poll(0x0004024C, 1, 30000)      (sleeping...)
poll(0x0004024C, 1, 30000)                      = 0
poll(0x0004024C, 1, 0)                          = 0
putmsg(4, 0xEFFFF6A4, 0xEFFFF7BC, 0)            = 0
poll(0x0004024C, 1, 14997)      (sleeping...)
poll(0x0004024C, 1, 14997)                      = 0
poll(0x0004024C, 1, 1)                          = 0
sigprocmask(SIG_SETMASK, 0xEFFFF768, 0xEFFFF758) = 0
close(4)                                        = 0
sigprocmask(SIG_SETMASK, 0xEFFFF758, 0x00000000) = 0
read(3, 0x00023E7C, 512)                        = 0
llseek(3, 0, SEEK_CUR)                          = 983
close(3)                                        = 0
llseek(0, 0, SEEK_CUR)                          = 254

i don't know if anyone can make heads or tails of this, but if so, please let
me know, or let me know what other info i need to get you.  this is more of
a curiosity than anything else, but it just started out of the blue not too
long ago, and i wonder what i did to change things. (nothing that i can think
of off hand)

both machines are sparc20s running Solaris 8 that was just installed yesterday
(new drives, woo!)

although, now that i think about it, doesn't quota ignore directories that are
NFS mounted that don't have quotas on?  that may be why it's not hanging on the
other machine since my home directory is NFS mounted.  i could be wrong, just
a brain fart before i finished my mail.


More information about the rescue mailing list