user@lists.sourceforge.net
Date: 2012-02-08 06:20
Subject:Re: [xcat-user] apache2 mod_cache, mod_file_cache,
mod_disk_cache
Hi Jonathan,
sounds brilliant.
Are you willing to share what you did for mod_disk_cache in terms of
configurations?
Thanks!
On 07.02.2012 18:05,
ever, and fail
> in postscripts.
>
> dnshandler is set to ddns in the site table and, again, this was all working
> yesterday.
>
> - jonathan
>
> - Original Message -
> From: "Jonathan Dye"
> To: "xCAT Users Mailing list"
> Sent: Tuesday, February 7, 2
site table and, again, this was all working
yesterday.
- jonathan
- Original Message -
From: "Jonathan Dye"
To: "xCAT Users Mailing list"
Sent: Tuesday, February 7, 2012 11:41:23 AM
Subject: [xcat-user] ddns update problems
suddenly DDNS updates are not happening
suddenly DDNS updates are not happening. the update key is set to the same
thing in both dhcp and named. i've tried cleaning the dhcp and dns state up to
the best of my knowledge. manual updates with that key work in the nsupdate
utility. i tried watching what happened when i update a lease
the gpxe config file generated by xnba.pm contains a line like:
imgargs kernel ... BOOTIF=01-${netX/machyp}
for some reason the diskless machines i'm booting don't expand ${netX/machyp}
and the kernel gets BOOTIF=01-
i've tried generating a hyphen-delimited mac for the boot interface and puttin
so, funny you ask, last week i ran into this problem and i tried using both
mod_mem_cache and mod_disk_cache. i set noderes.nfsserver to separate apache
servers configured to reverse proxy and cache (alternately i could have changed
the kickstart templates to use a different table entry and cha
is nodeset xxx install for esxi4.1 expected to work? i had to patch it to get
it to make an honest try (i think the esxi5 patches broke it), but the
hypervisors i try installing hang after all the files are transferred, possibly
during some sort of post install script but i haven't looked into
when i'm trying to boot a packed stateless image with 2.6 i'm getting this
error in the booting guest:
dracut: FATAL: No or empty root= argument
does anybody happen to know how i might be misconfiguring xcat that causes this
before i go try to trace through the dracut initrd image?
- jonathan
starting with the last bit, because i probably should have said this at the
outset:
> Currently you
> can't block ipmi.pm from getting 'rpower' if the user says
> nodehm.power=ipmi, but then I'm hard pressed to see a scenario where you'd
> have to do that.
i'm about done writing a simulation pl
- Original Message -----
From: "Jonathan Dye"
To: "xCAT Users Mailing list"
Sent: Tuesday, October 11, 2011 2:27:38 PM
Subject: [xcat-user] plugin_command and dispatch_request
are there places where we really want xcat to fulfill a single request with two
different plugins?
the plu
are there places where we really want xcat to fulfill a single request with two
different plugins?
the plugin_command code in xcatd and the copy of it in Client.pm will both
route requests to every plugin that makes it into the handler_hash for that
request's command. this is independent of th
what's broken about it?
- Original Message -
From: "Luis Miguel Silva"
To: "xCAT Users Mailing list"
Sent: Sunday, July 24, 2011 9:39:03 PM
Subject: [xcat-user] patched hpilo module (the one that ships with xCAT does
not seem to work)
Dear all,
About a year ago we, at Adaptive Comp
to created by Moab",
[root@xcat ~]# clonevm clonevm01 -b vmmaster01
clonevm01: Successfully cloned from vmmaster01
[root@xcat ~]#
This time it seemed to work!
I'm not very happy with how it works though... :o)
I understand it is necessary to specify the hypervisor where the VM
should be pla
you don't specify a full path for a clone, and you have the arguments backward.
you'll have to make a new VM since it rebased vm5 to the vm master called
'nfs_+JQ-2f+JQ-2f10.0.2.31+JQ-2fvms+JQ-2fshar'. then, you'll make a new master
like this:
clonevm -t [new_stateful_vm] vmmaster3
and you u
so, that's yet another problem. resolv.conf doesn't get written correctly. i
see where the initrd is supposed to be copying both the resolv.conf and the
lease file from the stateless environment, but it only copies the lease file.
when i manually fixed resolv.conf in the guest, the retries fr
the real question in that environment is actually why updateflag.awk is
failing. once you remove/background those 'tee's it just hangs on updateflag,
which seems to retry infinitely.
it seems to be connecting to the right port from the booting guest. i wasn't
very successful in trying to att
jarrod will have to correct me if i'm wrong, but i think the old makedns took a
network address and subnet mask as a sub-argument to -n. it was not like the
-n flag for makedhcp which writes out a fresh dhcpd.conf.
the new makedns -n does things along similar lines to makedhcp -n-- it writes
f
there was a version of ddns.pm that couldn't handle fake TLDs like that where
there was a TLD and no domain name (because of a bug in Net::DNS when the
domain didn't contain a dot, i think). the fix was a new ddns.pm, making sure
that resolv.conf pointed to the local machine, and restarting the
18 matches
Mail list logo
