[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] Condor Resourcelist



Would be nice if there was a deamon which polls ALL the information about the pool -- machines/resources, jobs (idle/held/running, etc...), accounting groups, priorities, etc..etc.. all in 1 central location. 

Presently, the accounting groups and priorities are with the negotiator. Jobs are with schedd. Machine resources are with collector, etc...




On Tue, Feb 22, 2011 at 5:19 PM, Ian Chesal <ichesal@xxxxxxxxxxxxxxxxxx> wrote:
The easiest route to go, to get things working in a test environment, is to accept all incoming requests at the negotiator and turn off any authentication methods in your cluster. If you're on a trusted network this usually isn't a big deal. If you're not, well, tread cautiously here.

Set all your clients to use:

SEC_DEFAULT_AUTHENTICATION_METHODS = CLAIMTOBE
SEC_CLIENT_AUTHENTICATION_METHODS = CLAIMTOBE


And make sure the permissions for actions look like the defaults:

HOSTALLOW_ADMINISTRATOR = $(CONDOR_HOST)
HOSTALLOW_OWNER = $(FULL_HOSTNAME), $(HOSTALLOW_ADMINISTRATOR)
HOSTALLOW_READ = *
HOSTALLOW_WRITE = *
HOSTALLOW_NEGOTIATOR = $(COLLECTOR_HOST)
HOSTALLOW_NEGOTIATOR_SCHEDD = $(COLLECTOR_HOST), $(FLOCK_NEGOTIATOR_HOSTS)
HOSTALLOW_WRITE_COLLECTOR = $(HOSTALLOW_WRITE), $(FLOCK_FROM)
HOSTALLOW_WRITE_STARTD    = $(HOSTALLOW_WRITE), $(FLOCK_FROM)
HOSTALLOW_READ_COLLECTOR  = $(HOSTALLOW_READ), $(FLOCK_FROM)
HOSTALLOW_READ_STARTD     = $(HOSTALLOW_READ), $(FLOCK_FROM)
HOSTALLOW_CLIENT = *


Once you have your pool functioning you can experiment with locking things down and different authentication methods to secure it.

Regards,
- Ian

On Tuesday, February 22, 2011 at 5:07 PM, Fritz, Eric R [AN S] wrote:

I checked the config files and the nodes are reporting to the host and when I checked the log file I have the following messages

02/22 14:46:03 PERMISSION DENIED to unauthenticated user from host * for command 0

and it has the same message for commands 1 and 2 as well. I have set the host config to allow for all machines on the network, so I am unsure of why they do not have permission.

-Eric

Eric Fritz
Assistant Scientist II
Iowa State University
2255G Kildee Hall
Ames, IA 50014
________________________________________
From: condor-users-bounces@xxxxxxxxxxx [condor-users-bounces@xxxxxxxxxxx] On Behalf Of Ian Chesal [ichesal@xxxxxxxxxxxxxxxxxx]
Sent: Tuesday, February 22, 2011 3:44 PM
To: Condor-Users Mail List
Subject: Re: [Condor-users] Condor Resourcelist

Okay, that means your other machines aren't reporting in to your Condor collector instance. You need to figure out why they aren't reporting in. Here's the list of things to start checking to see if it's...:

1. Configuration related. The other nodes aren't being told via the CONDOR_HOST/COLLECTOR_HOST macros where to find your central manger machine. See: http://www.cs.wisc.edu/condor/manual/v7.4/3_3Configuration.html#14219 -- check the condor_config and condor_config.local files on your execute nodes to make sure they're reporting in to your central manager.

<http://www.cs.wisc.edu/condor/manual/v7.4/3_3Configuration.html#14219>2. Security setting related. The other nodes are being denied access to report their status to the central manager. Check your CollectorLog file and see if there are write-denied messages in the log file.

3. Network related. Your other machines aren't connected to the same network as your central manager (this one sounds less plausible since you mentioned the network setup was sane already).

Regards,
- Ian

--
Ian Chesal
ichesal@xxxxxxxxxxxxxxxxxx

On Tuesday, February 22, 2011 at 4:39 PM, Fritz, Eric R [AN S] wrote:

It shows only the host computer.

-Eric


Eric Fritz
Assistant Scientist II
Iowa State University
2255G Kildee Hall
Ames, IA 50014
________________________________________
From: condor-users-bounces@xxxxxxxxxxx<mailto:condor-users-bounces@xxxxxxxxxxx> [condor-users-bounces@xxxxxxxxxxx<mailto:condor-users-bounces@xxxxxxxxxxx>] On Behalf Of Ian Chesal [ichesal@xxxxxxxxxxxxxxxxxx<mailto:ichesal@xxxxxxxxxxxxxxxxxx>]
Sent: Tuesday, February 22, 2011 3:26 PM
To: Condor-Users Mail List
Subject: Re: [Condor-users] Condor Resourcelist

What does:

condor_status -any

Show?

Regards,
- Ian

--
Ian Chesal
ichesal@xxxxxxxxxxxxxxxxxx<mailto:ichesal@xxxxxxxxxxxxxxxxxx>

On Tuesday, February 22, 2011 at 4:19 PM, Fritz, Eric R [AN S] wrote:

I have tried the condor_status command as well and this only shows the master computer and none of the computers that are supposed to be nodes, but when I do condor_findhost on a node computer it finds the correct host. I am unsure as to what the problem may be. It seems like the nodes can see the host but the host can't see the nodes except for itself.

-Eric


Eric Fritz
Assistant Scientist II
Iowa State University
2255G Kildee Hall
Ames, IA 50014
________________________________________
From: condor-users-bounces@xxxxxxxxxxx<mailto:condor-users-bounces@xxxxxxxxxxx> [condor-users-bounces@xxxxxxxxxxx<mailto:condor-users-bounces@xxxxxxxxxxx>] On Behalf Of Ian Chesal [ichesal@xxxxxxxxxxxxxxxxxx<mailto:ichesal@xxxxxxxxxxxxxxxxxx>]
Sent: Tuesday, February 22, 2011 11:51 AM
To: Condor-Users Mail List
Subject: Re: [Condor-users] Condor Resourcelist

Hi Eric,

Are you running a CondorView server in your pool? That's a somewhat old technology and it's not set up by default when you install Condor.

Are you mistaking condor_stats for the condor_status command? The condor_status command shows you detailed information about the machines in your pool.

Regards,
- Ian

--
Ian Chesal
ichesal@xxxxxxxxxxxxxxxxxx<mailto:ichesal@xxxxxxxxxxxxxxxxxx>

On Tuesday, February 22, 2011 at 12:47 PM, Fritz, Eric R [AN S] wrote:

Hi all,

I am trying to get Condor set up on a group of computers and was having what were thought to be network issues but the computers are now on a network with no restrictions in order to get full communication and when I use the command:

condor_stats -resourcelist

I get back the following message:

failed to receive data from the CondorView server computer_name

No Data.

Does anyone know what the possible problem is? I have gone through the configuration and tried all suggestions I could find to make the master see the nodes and nothing is working. The nodes can see the master but the master only sees its processors as nodes and none of the other computers and when I tried the above command that was the output i got back.


-Eric Fritz


Eric Fritz
Assistant Scientist II
Iowa State University
2255G Kildee Hall
Ames, IA 50014
_______________________________________________
Condor-users mailing list
To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx<mailto:condor-users-request@xxxxxxxxxxx> with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/condor-users

The archives can be found at:
https://lists.cs.wisc.edu/archive/condor-users/

_______________________________________________
Condor-users mailing list
To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx<mailto:condor-users-request@xxxxxxxxxxx> with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/condor-users

The archives can be found at:
https://lists.cs.wisc.edu/archive/condor-users/

_______________________________________________
Condor-users mailing list
To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx<mailto:condor-users-request@xxxxxxxxxxx> with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/condor-users

The archives can be found at:
https://lists.cs.wisc.edu/archive/condor-users/

_______________________________________________
Condor-users mailing list
To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/condor-users

The archives can be found at:
https://lists.cs.wisc.edu/archive/condor-users/


_______________________________________________
Condor-users mailing list
To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/condor-users

The archives can be found at:
https://lists.cs.wisc.edu/archive/condor-users/




--
--- Get your facts first, then you can distort them as you please.--