headers
Robert Mulley | 1 May 2003 03:13
Picon

A quick question about X.

Hello,

Due to some ridiculous requirement of Oracle we need to have an x server
running on our server. Oracle apparently uses the fonts that the X server
provides. We do not have a graphics card in the machines, is there a way to
run the x server software on the machine without a card? We don't care about
seeing the output. Config Trucluster GS320's running 5.1A PK3.

Thanks,

Robert Mulley
System Administrator
The Travel Corporation Pty Ltd 
Travel House 
35 Grafton Street 
Woollahra NSW 2025 
(02) 9028 5585
Permalink | Reply | 0 comments |
headers
Robert Mulley | 1 May 2003 06:32
Picon

SUMMARY: A quick question about X.

Quick question and lots of Quick answers. Thanks to Sue, James, Charles, Tom
and Danielle. All of whom suggested I use either Xvfb or Xvnc. So I'll go
try to get Xvfb working as I'd prefer something that is delivered with the
OS rather than third party.

Also thanks to Charles for pointing out the the ridiculous requirement comes
from the JVM rather than Oracle. Still Oracle did choose the Java path.

Original Question:
Due to some ridiculous requirement of Oracle we need to have an x server
running on our server. Oracle apparently uses the fonts that the X server
provides. We do not have a graphics card in the machines, is there a way to
run the x server software on the machine without a card? We don't care about
seeing the output. Config Trucluster GS320's running 5.1A PK3.

Regards,

Robert Mulley
System Administrator
The Travel Corporation Pty Ltd 
Travel House 
35 Grafton Street 
Woollahra NSW 2025 
(02) 9028 5585
Permalink | Reply | 0 comments |
headers
Chris Bryant | 1 May 2003 17:46

SUMMARY: C2 Enhanced Security and null passwords

Thanks to Pat O'Brien and Martin Moore for their responses.  When I
configured the /etc/auth/system/default file to take null passwords and then
turn on C2 security the null passwords remained null.

Thanks again,

Chris

-----Original Message-----
From: Chris Bryant 
Sent: Tuesday, April 22, 2003 8:58 AM
To: Tru64 Unix (tru64-unix-managers <at> ornl.gov)
Subject: C2 Enhanced Security and null passwords

Admins,
	
	I am in the planning process to convert a Tru64 UNIX 5.1 system from
BASE security to ENHANCED security. Now, I have a lot of accounts that do
not have passwords associated with them and I know that ENHANCED security
supports null passwords.  My question is after I convert to enhanced
security will I have to remove the password field or will it just keep the
passwords null?  I will summarize.

TIA,

Chris Bryant
Unix Administrator
Dollar Thrifty Automotive Group
(918) 669-3213
(Continue reading)

Permalink | Reply | 0 comments |
headers
Cohen, Andy | 1 May 2003 18:25
Favicon

SUMMARY: disable desktop/fix video?

I think I fixed it.  A simple power-cycling of the box did it.  I made the
mistake of not powering down and back on after installing the firmware.  I
think this happened for 2 reasons: 1) the documentation I was using didn't
mention it and 2) the firmware actually didn't change versions.  I then went
on to read the installation guide and it says to power down after upgrading
firmware so I did that and now CDE is fine.

I did get 2 very helpful emails from Rogier Sars that I thought I should
pass along:

===========
It could be that you are using a seperate gaphics card installed in one of
the expansion slots. If your model of 1000A also has an onboard graphics
chip this will cause a conflict. Check the boot messages to confirm this. If
this is the case you should disable the onboard graphics chip. This is done
by going into ARC and run the ECU from there. ECU is the EISA Configuration
Utillity and should have come with your system.
============

and (after I sent him the boot messages from /var/adm/messages):

============
I don't see any other VGA chipset in the boot messages besides the Cirrus
Logic 5428. According to my tech docs the CL 5428 is the onboard graphics
chip.

When I got the Alphaserver 1000A I have at home I had exactly the same
problem as you describe. The CDE was in black and white and had lines
running through it. After a while it would crash back to the command prompt.
In my case it turned at that I had to disable the onboard graphics chip
(Continue reading)

Permalink | Reply | 0 comments |
headers
Senulis, Joseph A | 1 May 2003 20:30
Picon
Picon

Upgrade from HSZ70 to HSZ80

Hi,
     I thought that I had seen this addressed previously, but I couldn't
find it in the archives.  We are currently running T64 5.1 and plan to
upgrade to 5.1B.  One of our systems, standalone, not clustered, has a BA370
with a single HSZ70 controller and all RAID sets either RAID-1 or RAID-0+1.
Since the HSZ70 will not be supported under 5.1B, the simplest solution
seems to be to just swap the HSZ70 with the HSZ80.  Is it really that easy?
I haven't placed a service call with HP to get the answer but the sales folk
are hazy about it.  Doing a full backup/restore would work but I would like
to have less down-time.  I could probably temporarily borrow an extra host
adapter to allow the HSZ70 and HSZ80 to be used simultaneously during a
cut-over if that would help.  Thanks and I will summarize.
--Joe
Permalink | Reply | 0 comments |
headers
Senulis, Joseph A | 1 May 2003 21:55
Picon
Picon

SUMMARY: Upgrade from HSZ70 to HSZ80

Thanks to Dr. Thomas Blinn, Toni Harbaugh-Blackford, Alan Rollow and Bruce
D. Hines for the speedy reply.

>From Bruce Hines:
---
Steps to swap HSZ70 to HSZ80
1. Backup all data on HSZ70
2. Capture all HSZ70 configuration information:
	a. show this_controller full
	b. show disks
	c. show storagesets full
	d. show devices full
3. Shutdown host and HSZ70
4. HP does their thing
5. run config (to add all devices it finds, like DISK10000, etc)
6. add mirrors, add stripes, add spares  (DO NOT INITIALIZE ANYTHING)
7. add units (I would use same names for all containers and units)
8. boot host

I would detail out the steps with each container name with commands (makes
it easier to cut-n-paste when you get ready to do it) then run your plan by
HP for proof of concept.
---
     Steps 1. and 2. are already done as part of our disaster recovery
procedures and the rest isn't all that painful.  If push comes to shove,
that will be the way that we go.  However, hope comes from Dr. Thomas Blinn:
---
What you propose SHOULD work.  However, you should also ask when there
might be HSZ70 support under V5.1B.  My understanding is that we have
committed to provide such support precisely because of situations like
(Continue reading)

Permalink | Reply | 0 comments |
headers
Ballowe, Charles | 2 May 2003 00:22

Performance questions

I have a GS-80 (clustered actually) running Tru64 5.1A PK 4, serving Oracle
8i
databases. I have some questions that some people might have tested or know
answers
to.

For reference, the system has 4CPUs and 6G of RAM on 3 cards all in one QBB.
If I remember correctly, this puts the memory at 4way interleaving.

I've recently been seeing somewhat higher percentages of the time reported
as
system time (both vmstat and collect). Is there any way to see what portions
of the kernel are being used during this time? (Higher percentages is
sometimes
20% during heavy load)

I've also been seeing load averages reported near 4 or higher under some
peak load
conditions, but seeing a fair chunk (20%+) marked idle. With a load of 4 and
4 CPUs
there should always be something ready to run - is there anything that might
explain
this?

I'm also curious if there's a way to see how much time the system is
spending 
fetching data from memory, or L2 cache into the L1 cache as well as cache
hit/miss
ratios. I'm wondering if adding a 4th memory board with 2G, which I believe
would
(Continue reading)

Permalink | Reply | 0 comments |
headers
Rakowski | 2 May 2003 01:26
Picon

Need tyo move unraided disks

I have a client whose data disks were initially setup as JBOD.  Since being
setup some years ago, they have added disks, and put the new disks in raid
sets.  So now they have 2 raid sets (presented as 2 disks) and 6 unprotected
9GB disks, all of which are being used.

Fortunately, the OS is not on the disks in question.  Unfortunately, the
client has added and extended volumes over the years so that the file
systems and raw database devices are intertwined through the 6 unprotected
disks and the 2 raidsets.   The raw devices won't be an issue after next
week as the client is converting their database and I can get rid of them.
The file systems will still be there though.

Any thoughts on the best way to put these 6 unprotected disks into a raid
set without wiping out the data and restoring (before one of them dies and I
have to do a complete restore anyway...)?  The system is an ES40 running
5.1A with an HSZ70.  Another note is that although the OS is not in LSM, the
entire app and database are in rootdg...

Any help would be appreciated.

Rob
Permalink | Reply | 0 comments |
headers
Parkin Frank - fparki | 2 May 2003 11:26
Picon

SUMMARY:Samba and TruClusters


Original Post:
--------------

> Has anyone got Samba working as a CAA whilst using the cluster alias?
>
> TruCluster v5.1a
> Samba v2.2.8a
>
> Many thanks
Frank

Summary:
--------

Thanks for responses from Chris Cooke, Emanuele Lombardi, Massimiliano
Balladelli, Jim Lola, Blake Roberts, Wolfgang Rupp and Wolfram Klaus.

The CAA part was relatively easy. Just take an existing *.scr and *.cap file
and Samba fails-over/relocates without any problems. 

To get Samba to work with the cluster alias requires the following lines in
smb.conf. 

netbios name = <cluster alias name>
interfaces = <cluster alias IP address>

>From samba.org...

Netbios name
(Continue reading)

Permalink | Reply | 0 comments |
headers
Rezk Mekhael | 2 May 2003 18:02

ftp with some user

Hi managers,

I am getting this message with only one user when he tray to ftp 

"ftp: ftp/tcp: unknown service"
and same thing when he tray to do it from /usr/bin/ftp

any idea?

--
Sincerely,
Rezk Mekhael
Manager, Internet Systems and Services
ISS Department
California Lutheran University
60 W. Olsen Rd.
Thousand Oaks, CA 91361
Office  : (805) 493-3947
Pager  : (805) 230-4926
Cell     : (818) 634-9410
Fax     : (805) 493-3842
E-mail : Rezk <at> clunet.edu
 

*** Sent Mail scanned for known viruses by CLUnet ***
Permalink | Reply | 0 comments |

Gmane