|
Bug ID
|
4530367
|
|
Synopsis
|
After retry timeout - nss_search() no longer retries lookups
|
|
State
|
10-Fix Delivered (Fix available in build)
|
|
Category:Subcategory
|
network:nsswitch
|
|
Keywords
|
nss_search | res_gethostbyname | resolv.conf | retrans | retry | retry_test | sunalert-submit-availability
|
|
Responsible Engineer
|
Stacey Marshall
|
|
Reported Against
|
5.8
|
|
Duplicate Of
|
|
|
Introduced In
|
solaris_8
|
|
Commit to Fix
|
s10_14
|
|
Fixed In
|
s10_14
|
|
Release Fixed
|
solaris_10(s10_14)
,
solaris_9u3(s9u3_06) (Bug ID:2049273)
|
|
Related Bugs
|
6275116
,
4260677
|
|
Submit Date
|
21-November-2001
|
|
Last Update Date
|
16-August-2007
|
|
Description
|
When using the resolver from the nsswitch, timeouts can occur due to non-responding or poorly configured servers. When this happens there
is an apparent state change which prevents the nss_search() loop from
retrying destination. This state appears to persist unless application
is restarted and a new handle into the nsswitch is generated.
|
|
Work Around
|
1) Improve name serer services RAS
2) Restart application.
3) Increase retrans to a sufficient level in order to guarante that a slow
servers response is seen; refer to resolv.conf(1m) man page for details.
This may introduce an unacceptable latency into the nsswitch.
|
|
Comments
|
N/A
|