[prev in list] [next in list] [prev in thread] [next in thread]
List: npaci-rocks-discussion
Subject: Re: [Rocks-Discuss] problems in setting up compute nodes with 5.4.3
From: "Doll, Margaret Ann" <margaret_doll () brown ! edu>
Date: 2011-09-23 16:07:52
Message-ID: CAKhbxNngDDcwcE=to6QYZrGNQ1UOnM9-_3=WZXArh+0BAQoi7w () mail ! gmail ! com
[Download RAW message or body]
Thank you again. I am sorry that I am have pestered you so much. I will
keep notes of all that I have learned and pass them on to the next manager
that I see.
On Fri, Sep 23, 2011 at 12:02 PM, Philip Papadopoulos <
philip.papadopoulos@gmail.com> wrote:
> On Fri, Sep 23, 2011 at 7:00 AM, Doll, Margaret Ann <
> margaret_doll@brown.edu
> > wrote:
>
> > Thanks. In the first email the command looked like a two line command.
> >
> > I used
> >
> > [root@ted db]# cd /var/db
> > [root@ted db]# ls
> > Makefile mysql-backup-cluster nscd RCS
> > [root@ted db]# rm mysql-backup-cluster
> > rm: remove regular file `mysql-backup-cluster'? y
> > [root@ted db]# co -d 2001-09-17 10:00:00:00-00 mysql-backup-cluster
> > co: RCS/2001-09-17,v: No such file or directory
> > co: RCS/10:00:00:00-00,v: No such file or directory
> > RCS/mysql-backup-cluster,v --> mysql-backup-cluster
> > revision 1.3
> > done
> > [root@ted db]# /opt/rocks/bin/mysql
> > --defaults-extra-file=/root/.rocks.my.cnf -u root cluster <
> > mysql-backup-cluster
> > [root@ted db]# rocks sync config
> > [root@ted db]# rocks list host
> >
> > Gives me the complete list of hosts again, including the double
> > compute-0-2s. I built the system on the morning of the 17th. I didn't
> > have
> > all the compute nodes added until early evening.
> >
> > Is there a way to emplty out all the compute nodes from the mysql
> database
> > and then start the re-install of the power units and compute nodes?
> >
> > What is the applicance name for the power distribution units when you use
> > rocks set host applicance
> >
>
> If you do a rocks list host you will see a category called "membership".
>
> You can also do "rocks list membership" to show the various memberships.
>
> [ Aside: A membership is used internally and
> is an (Appliance, Distribution) pair. Nearly every Rocks cluster has only
> one distribution (rocks-dist), but can support multiple (with some
> small amount of additional SQL work). ]
>
> So, you would want to do the following
>
> # rocks set host membership power-0-2 "Power Distribution Unit"
>
> Do that for all of your "power management" units. That allows Rocks to
> know
> about the mac addresses and IP addresses, but it won't try
> to manage these appliances in any way
>
> After that, sync -should- work.
>
> -P
>
>
>
> > Thanks so much for your help
> >
> > On Fri, Sep 23, 2011 at 9:49 AM, Philip Papadopoulos <
> > philip.papadopoulos@gmail.com> wrote:
> >
> > > you have to feed the dumped database on the command line.
> > > I think email broke the command across two lines ... the
> > > "/opt/rocks/bin/mysql --defaults-extra-file=/root/.rocks.my.cnf -u root
> > > cluster < mysql-backup-cluster"
> > >
> > > (remove the quotes, but it is all one command)
> > >
> > > -P
> > >
> > > On Fri, Sep 23, 2011 at 5:51 AM, Doll, Margaret Ann <
> > > margaret_doll@brown.edu
> > > > wrote:
> > >
> > > > On Wed, Sep 21, 2011 at 8:11 PM, Philip Papadopoulos <
> > > > philip.papadopoulos@gmail.com> wrote:
> > > >
> > > > > On Wed, Sep 21, 2011 at 12:48 PM, Doll, Margaret Ann <
> > > > > margaret_doll@brown.edu> wrote:
> > > > >
> > > > > > How do I revert back to a version of the DB that has no
> duplicate?
> > > > > > How do I see what versions are in /var/db/mysql-backup-cluster?
> > > > > >
> > > > > cd /var/db
> > > > > rc2slog mysql-backup-cluster
> > > > >
> > > > > You can get the first version of your database
> > > > >
> > > > > # cd /var/db
> > > > > # rm mysql-backup-cluster
> > > > > # co -r 1.1 mysql-backup-cluster
> > > > >
> > > > > (Note there are other options for getting versions of rcs'ed files,
> > you
> > > > ask
> > > > > for by date, too. so the "co" man page)
> > > > >
> > > > > Then restore that version of the database with
> > > > > # /opt/rocks/bin/mysql --defaults-extra-file=/root/.rocks.my.cnf -u
> > > root
> > > > > cluster < mysql-backup-cluster
> > > > > # rocks sync config
> > > > >
> > > >
> > > > Thio did not work:
> > > >
> > > > mysql> cluster < mysql-backup-cluster;
> > > > ERROR 1064 (42000): You have an error in your SQL syntax; check the
> > > manual
> > > > that corresponds to your MySQL server version for th
> > > > e right syntax to use near 'cluster < mysql-backup-cluster' at line 1
> > > >
> > > >
> > > > rocks remove compute
> > > >
> > > > did nothing because I have a couple of compute nodes that are listed
> > > twice.
> > > >
> > > > rocks remove compute-0-40
> > > > rocks sync config
> > > >
> > > > Did not remove compute-0-40
> > > >
> > > > insert-ethers --remove compute-0-40
> > > > rocks sync config
> > > >
> > > > did not remove compute-0-40
> > > >
> > > > >
> > > > > -P
> > > > >
> > > > > >
> > > > > > Thanks to everyone that is giving me information.
> > > > > >
> > > > > > On Wed, Sep 21, 2011 at 3:37 PM, Philip Papadopoulos <
> > > > > > philip.papadopoulos@gmail.com> wrote:
> > > > > >
> > > > > > > you will need to revert to a version of the DB that has no
> > > duplicate
> > > > > > > entries.
> > > > > > >
> > > > > > > what versions are available in the RCS'ed
> > > > /var/db/mysql-backup-cluster
> > > > > > (See
> > > > > > > previous email)
> > > > > > >
> > > > > > >
> > > > > > > -P
> > > > > > >
> > > > > > > On Wed, Sep 21, 2011 at 12:21 PM, Doll, Margaret Ann <
> > > > > > > margaret_doll@brown.edu> wrote:
> > > > > > >
> > > > > > > > "rocks remove host compute-0-2" will not remove compute-0-2
> > > because
> > > > > > there
> > > > > > > > are two of them listed.
> > > > > > > >
> > > > > > > > On Wed, Sep 21, 2011 at 2:11 PM, Ian Kaufman <
> > > > ikaufman@soe.ucsd.edu>
> > > > > > > > wrote:
> > > > > > > >
> > > > > > > > > I would try removing all nodes and starting over.
> > > > > > > > >
> > > > > > > > > I believe you need to use:
> > > > > > > > >
> > > > > > > > > rocks remove host <hostname>
> > > > > > > > >
> > > > > > > > > Ian
> > > > > > > > >
> > > > > > > > > On Mon, Sep 19, 2011 at 6:35 AM, Doll, Margaret Ann
> > > > > > > > > <margaret_doll@brown.edu> wrote:
> > > > > > > > > > I have a problem with the compute nodes that I installed
> > with
> > > > > Rocks
> > > > > > > > > 5.4.3.
> > > > > > > > > > The compute nodes are installed, but what used to be
> > > > maintenance
> > > > > > > nodes
> > > > > > > > > are
> > > > > > > > > > also installed as compute nodes. I used insert-ethers
> to
> > > > > install
> > > > > > > the
> > > > > > > > > > compute nodes. For each installation I plugged in new
> > blade
> > > at
> > > > a
> > > > > > > time
> > > > > > > > > and
> > > > > > > > > > started up one compute node at a time. The first
> > > installation
> > > > on
> > > > > a
> > > > > > > > > blade,
> > > > > > > > > > installed what were the maintenance nodes on the old
> system
> > > as
> > > > > > > compute
> > > > > > > > > nodes
> > > > > > > > > > as well as installing the real compute node. The second
> > > > > > installation
> > > > > > > > on
> > > > > > > > > the
> > > > > > > > > > blade picked up the second compute node.
> > > > > > > > > >
> > > > > > > > > > from rocks list host boot
> > > > > > > > > >
> > > > > > > > > > compute-0-34: install
> > > > > > > > > > compute-0-35: install
> > > > > > > > > > compute-0-36: os
> > > > > > > > > > compute-0-37: os
> > > > > > > > > >
> > > > > > > > > > I noticed with insert-ethers, maintenance nodes are no
> > longer
> > > > > > listed.
> > > > > > > > > What
> > > > > > > > > > are these second enties now called. Each of my compute
> > nodes
> > > > has
> > > > > > two
> > > > > > > > > > ethernet connections.
> > > > > > > > > >
> > > > > > > > > > What should these previous maintenance nodes be called?
> > How
> > > > do
> > > > > I
> > > > > > > > change
> > > > > > > > > > them from compute nodes?
> > > > > > > > > >
> > > > > > > > > > compute-0-34: private ----- 00:30:48:99:59:67
> 192.168.6.223
> > > > > > > > 255.255.255.0
> > > > > > > > > > ------ compute-0-34 ---- ------- -------
> > > > > > > > > > compute-0-35: private ----- 00:30:48:99:59:79
> 192.168.6.222
> > > > > > > > 255.255.255.0
> > > > > > > > > > ------ compute-0-35 ---- ------- -------
> > > > > > > > > > compute-0-36: ------- eth1 00:30:48:7d:d2:c1
> -------------
> > > > > > > > > ---------------
> > > > > > > > > > ------ ------------ ---- ------- -------
> > > > > > > > > > compute-0-36: private eth0 00:30:48:7d:d2:c0
> 192.168.6.221
> > > > > > > > 255.255.255.0
> > > > > > > > > > ------ compute-0-36 ---- ------- -------
> > > > > > > > > > compute-0-37: ------- eth1 00:30:48:7d:d2:c5
> -------------
> > > > > > > > > ---------------
> > > > > > > > > > ------ ------------ ---- ------- -------
> > > > > > > > > > compute-0-37: private eth0 00:30:48:7d:d2:c4
> 192.168.6.220
> > > > > > > > 255.255.255.0
> > > > > > > > > > ------ compute-0-37 ---- ------- -------
> > > > > > > > > >
> > > > > > > > > > I also tried to get the numbering correct by using "rocks
> > add
> > > > > host"
> > > > > > > I
> > > > > > > > > used
> > > > > > > > > > this to add compute-0-1, compute-0-2 and compute-0-3.
> > When
> > > I
> > > > > then
> > > > > > > > > install
> > > > > > > > > > compute-0-1, I got two entries
> > > > > > > > > >
> > > > > > > > > > compute-0-1: ------- eth0 00:30:48:79:c8:db
> -------------
> > > > > > > > > ---------------
> > > > > > > > > > ------ ------------ ---- ------- -------
> > > > > > > > > > compute-0-1: ------- eth0 00:30:48:79:c8:db
> -------------
> > > > > > > > > ---------------
> > > > > > > > > > ------ ------------ ---- ------- -------
> > > > > > > > > > compute-0-1: private eth1 00:30:48:79:c8:da
> 192.168.6.252
> > > > > > > > 255.255.255.0
> > > > > > > > > > ------ compute-0-1 ---- ------- -------
> > > > > > > > > > compute-0-1: private eth1 00:30:48:79:c8:da
> 192.168.6.252
> > > > > > > > 255.255.255.0
> > > > > > > > > > ------ compute-0-1 ---- ------- -------
> > > > > > > > > >
> > > > > > > > > > but they both show CPUs as being 1 instead of 8
> > > > > > > > > >
> > > > > > > > > > compute-0-1: Compute 1 0 1 os install
> > > > > > > > > > compute-0-1: Compute 1 0 1 os install
> > > > > > > > > >
> > > > > > > > > > I can ssh into compute-0-1
> > > > > > > > > >
> > > > > > > > > > I have two entries for
> > > > > > > > > >
> > > > > > > > > > compute-0-2: Compute 1 0 2 os install
> > > > > > > > > > compute-0-2: Compute 1 0 2 os install
> > > > > > > > > > compute-0-3: Compute 1 0 3 os install
> > > > > > > > > > compute-0-3: Compute 1 0 3 os install
> > > > > > > > > >
> > > > > > > > > > for boot
> > > > > > > > > >
> > > > > > > > > > compute-0-2: os
> > > > > > > > > > compute-0-2: os
> > > > > > > > > > compute-0-3: -------
> > > > > > > > > > compute-0-3: -------
> > > > > > > > > >
> > > > > > > > > > compute-0-2: private ----- 00:30:48:8c:43:21
> 192.168.6.253
> > > > > > > > 255.255.255.0
> > > > > > > > > > ------ compute-0-2 ---- ------- -------
> > > > > > > > > > compute-0-2: private ----- 00:30:48:8c:43:21
> 192.168.6.253
> > > > > > > > 255.255.255.0
> > > > > > > > > > ------ compute-0-2 ---- ------- -------
> > > > > > > > > > compute-0-3: private ----- 00:30:48:8c:43:3c
> 192.168.6.251
> > > > > > > > 255.255.255.0
> > > > > > > > > > ------ compute-0-3 ---- ------- -------
> > > > > > > > > > compute-0-3: private ----- 00:30:48:8c:43:3c
> 192.168.6.251
> > > > > > > > 255.255.255.0
> > > > > > > > > > ------ compute-0-3 ---- ------- -------
> > > > > > > > > >
> > > > > > > > > > insert-ethers --remove=compute-0-2 does not remove
> > either
> > > > of
> > > > > > the
> > > > > > > > > > compute-0-2 entries. Both of these ethernet addresses
> > used
> > > to
> > > > > > > belong
> > > > > > > > to
> > > > > > > > > > maintenance nodes on my previous installation.
> > > > > > > > > >
> > > > > > > > > > How do I fixed this installation? If I have to
> reinstall
> > > the
> > > > > > > compute
> > > > > > > > > > nodes, how do I clear out the previous tables?
> > > > > > > > > >
> > > > > > > > > > Thanks for your help
> > > > > > > > > > -------------- next part --------------
> > > > > > > > > > An HTML attachment was scrubbed...
> > > > > > > > > > URL:
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
> https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/attachments/20110919/6b3e63ea/attachment.html
>
> > > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > --
> > > > > > > > > Ian Kaufman
> > > > > > > > > Research Systems Administrator
> > > > > > > > > UC San Diego, Jacobs School of Engineering ikaufman AT ucsd
> > DOT
> > > > edu
> > > > > > > > >
> > > > > > > > -------------- next part --------------
> > > > > > > > An HTML attachment was scrubbed...
> > > > > > > > URL:
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
> https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/attachments/20110921/a7842e27/attachment.html
>
> > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > --
> > > > > > > Philip Papadopoulos, PhD
> > > > > > > University of California, San Diego
> > > > > > > 858-822-3628 (Ofc)
> > > > > > > 619-331-2990 (Fax)
> > > > > > > -------------- next part --------------
> > > > > > > An HTML attachment was scrubbed...
> > > > > > > URL:
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
> https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/attachments/20110921/499e284f/attachment.html
>
> > > > > > >
> > > > > > -------------- next part --------------
> > > > > > An HTML attachment was scrubbed...
> > > > > > URL:
> > > > > >
> > > > >
> > > >
> > >
> >
> https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/attachments/20110921/d01c0630/attachment.html
>
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Philip Papadopoulos, PhD
> > > > > University of California, San Diego
> > > > > 858-822-3628 (Ofc)
> > > > > 619-331-2990 (Fax)
> > > > > -------------- next part --------------
> > > > > An HTML attachment was scrubbed...
> > > > > URL:
> > > > >
> > > >
> > >
> >
> https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/attachments/20110921/2907f4ad/attachment.html
>
> > > > >
> > > > -------------- next part --------------
> > > > An HTML attachment was scrubbed...
> > > > URL:
> > > >
> > >
> >
> https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/attachments/20110923/be4ad4a3/attachment.html
>
> > > >
> > > >
> > >
> > >
> > > --
> > > Philip Papadopoulos, PhD
> > > University of California, San Diego
> > > 858-822-3628 (Ofc)
> > > 619-331-2990 (Fax)
> > > -------------- next part --------------
> > > An HTML attachment was scrubbed...
> > > URL:
> > >
> >
> https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/attachments/20110923/0d648a99/attachment.html
>
> > >
> > -------------- next part --------------
> > An HTML attachment was scrubbed...
> > URL:
> >
> https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/attachments/20110923/385f4a8b/attachment.html
>
> >
> >
>
>
> --
> Philip Papadopoulos, PhD
> University of California, San Diego
> 858-822-3628 (Ofc)
> 619-331-2990 (Fax)
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL:
> https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/attachments/20110923/81bed339/attachment.html
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/attachments/20110923/3d88cf1c/attachment.html \
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic