Next: solaris9 - locale
From: Stephen Hahn on 8 Feb 2005 17:12
* Gary Mills <mills(a)cc.umanitoba.ca>:
> In <42067fc9$0$28980$e4fe514c(a)news.xs4all.nl> Casper H.S. Dik <Casper.Dik(a)Sun.COM> writes:
>
> >mills(a)cc.umanitoba.ca (Gary Mills) writes:
>
>
>>> Feb 5 20:05:05/4 ERROR: Restarter: Ignoring REMOVE_INSTANCE command for unknown service svc:/TEMP/site/svcadjust:default.
>>> Feb 5 20:05:15/8: Ignoring svc:/TEMP/network/dns/server:default because its enabled property is misconfigured.
>>> Feb 5 20:05:15/4 ERROR: Restarter: Ignoring REMOVE_INSTANCE command for unknown service svc:/TEMP/network/dns/server:default.
>>> Feb 5 20:06:40/1: restarting after interruption
>>> Feb 5 20:06:44/74: svc:/application/font/fc-cache:default: transient ctid expected to be 0 but was set to 39
>>> Feb 5 20:06:48/113 ERROR: svc:/site/zephyr-hm:default: Method "/lib/svc/method/site/zephyr-hm" failed with exit status 99.
>>> Feb 5 20:06:48/113: site/zephyr-hm:default failed
>
>
> How do I track down the problem? It was okay with b72.

The "TEMP" portion of the FMRIs allows us to associate these messages
as occurring as a portion of the manifest upgrades.

The svc.startd failure is happening because of incorrect caching by
the linnker for some unavailable lazily-loaded library dependencies.
You have a separate / and /usr and one or more services that run as
non-root, right?

Being tracked as

6215444 ld.so.1 caches "not there" lazy libraries, foils
svc.startd(1M)'s logic

If you saw this *without* split / and /usr, please let me know.

- Stephen

--
Stephen Hahn, PhD Solaris Kernel Development, Sun Microsystems
stephen.hahn(a)sun.com
From: Gary Mills on 8 Feb 2005 18:22
In <cubdfn$avl$1(a)news1nwk.SFbay.Sun.COM> Stephen Hahn <stephen.hahn(a)s-u-n.c-o-m> writes:

>* Gary Mills <mills(a)cc.umanitoba.ca>:
>> In <42067fc9$0$28980$e4fe514c(a)news.xs4all.nl> Casper H.S. Dik <Casper.Dik(a)Sun.COM> writes:
>>
>> >mills(a)cc.umanitoba.ca (Gary Mills) writes:
>>
>>>> Feb 5 20:06:40/1: restarting after interruption
>>
>> How do I track down the problem? It was okay with b72.

> The svc.startd failure is happening because of incorrect caching by
> the linnker for some unavailable lazily-loaded library dependencies.
> You have a separate / and /usr and one or more services that run as
> non-root, right?

Yes, /usr and / are separate partitions. Yes, big-brother runs
as daemon. That'll be the problem.

> Being tracked as

> 6215444 ld.so.1 caches "not there" lazy libraries, foils
> svc.startd(1M)'s logic

Excellent! Thanks.

--
-Gary Mills- -Unix Support- -U of M Academic Computing and Networking-
First  |  Prev  | 
Pages: 1 2
Next: solaris9 - locale