hi there,
I have running a nsd test server with about 100.000 zones as master and
one as slave in my test environment. The master works pretty good but
the slave for all these zone don't.
The startup takes about an hour in which the nsd answers to all requests
with servfail instead of refusing the connection. In this time it is
fetching all serial from the master according to some straces I made to
the master process. But the master process keeps consuming 100% CPU time
after that. Here is a short strace clip of what it is doing:
gettimeofday({1183623774, 898647}, NULL) = 0
pselect6(10, [9], [9], [], {0, 101353000}, {0, 8}) = 1 (out [9], left
{0, 101353000})
write(9, "\10\0\0\0\0\21\1\v\3\n\6\0\5domain1\3com\0", 23) = 23
gettimeofday({1183623774, 923962}, NULL) = 0
pselect6(10, [9], [9], [], {0, 76038000}, {0, 8}) = 1 (out [9], left {0,
76038000})
write(9, "\10\0\0\0\0\27\1\21\3\20\f\0\vdomain34\3com\0", 29) = 29
gettimeofday({1183623774, 949287}, NULL) = 0
pselect6(10, [9], [9], [], {0, 50713000}, {0, 8}) = 1 (out [9], left {0,
50713000})
write(9, "\10\0\0\0\0\27\1\21\3\20\f\0\vdomainxy\3com\0", 29) = 29
gettimeofday({1183623774, 974594}, NULL) = 0
pselect6(10, [9], [9], [], {0, 25406000}, {0, 8}) = 1 (out [9], left {0,
25406000})
write(9, "\10\0\0\0\0\34\1\26\3\25\21\0\20domsdy\3co"..., 34) = 34
gettimeofday({1183623774, 999916}, NULL) = 0
pselect6(10, [9], [9], [], {0, 84000}, {0, 8}) = 1 (out [9], left {0,
84000})
write(9, "\10\0\0\0\0\33\1\25\3\24\20\0\17asdddd23489jfkl\3com"...,
33) = 33
gettimeofday({1183623775, 25220}, NULL) = 0
It looks like it is nonstop trying to refresh zones.
Any suggestions ?
Jan
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 252 bytes
Desc: OpenPGP digital signature
URL:
<http://lists.nlnetlabs.nl/pipermail/nsd-users/attachments/20070705/adc10560/attachment.bin>