[gridway-user] Problem with gridway installation

Javier Fontan jfontan at gmail.com
Thu Nov 15 05:48:56 CST 2007


Hello,

Could you please send me the output of these two commands?

wsrf-query -a -z none -s
https://grid1ag.g1dominio.com:8443/wsrf/services/DefaultIndexService

wsrf-query -a -z none -s
https://grid2ag.g2dominio.com:8443/wsrf/services/DefaultIndexService


Bye

On 11/14/07, Eddy Diaz <eddydiaz at gmail.com> wrote:
> Thanks for your suggestions; I can fix the problem reinstalling gridway.
>
> But now I have a new problem.
>
> My configuration file
>
> IM_MAD = mds4:gw_im_mad_mds4_thr:-l host.list:gridftp:ws
>
> EM_MAD = ws:gw_em_mad_ws::rsl2
>
> TM_MAD = gridftp:gw_tm_mad_ftp:
>
>
>
> Now, my log file looks
>
>
> [scheduler at scheduler ~]$ cat /usr/local/gw/var/gwd.log
>
> Wed Nov 14 07:54:04 2007 [GW][I]:
> ---------------------------------------------------
>
> Wed Nov 14 07:54:04 2007 [GW][I]:                    gwd.conf
> values
>
> Wed Nov 14 07:54:04 2007 [GW][I]:
> ---------------------------------------------------
>
> Wed Nov 14 07:54:04 2007 [GW][I]:   Core configuration
> attributes
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     GWD_PORT                 : 6725
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     MAX_NUMBER_OF_CLIENTS    : 25
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     NUMBER_OF_ARRAYS         : 200
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     NUMBER_OF_JOBS           : 5000
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     NUMBER_OF_HOSTS          : 100
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     NUMBER_OF_USERS          : 30
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     SCHEDULING_INTERVAL      : 30
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     DISCOVERY_INTERVAL       : 900
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     MONITORING_INTERVAL      : 300
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     POLL_INTERVAL            : 180
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     MAX_ACTIVE_IM_QUERIES    : 10
>
> Wed Nov 14 07:54:04 2007 [GW][I]:   Information Manager MADs
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     MAD(0)  name  : mds4
>
> Wed Nov 14 07:54:04 2007 [GW][I]:         executable: gw_im_mad_mds4_thr
>
> Wed Nov 14 07:54:04 2007 [GW][I]:         argument  : -l host.list
>
> Wed Nov 14 07:54:04 2007 [GW][I]:         TM        : gridftp
>
> Wed Nov 14 07:54:04 2007 [GW][I]:         EM        : ws
>
> Wed Nov 14 07:54:04 2007 [GW][I]:   Transfer Manager MADs
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     MAD(0)  name  : gridftp
>
> Wed Nov 14 07:54:04 2007 [GW][I]:         executable: gw_tm_mad_ftp
>
> Wed Nov 14 07:54:04 2007 [GW][I]:         argument  :
>
> Wed Nov 14 07:54:04 2007 [GW][I]:   Execution Manager MADs
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     MAD(0)  name  : ws
>
> Wed Nov 14 07:54:04 2007 [GW][I]:         executable: gw_em_mad_ws
>
> Wed Nov 14 07:54:04 2007 [GW][I]:         argument  :
>
> Wed Nov 14 07:54:04 2007 [GW][I]:         rsl mode  : rsl2
>
> Wed Nov 14 07:54:04 2007 [GW][I]:   Dispatch Manager Scheduler
>
> Wed Nov 14 07:54:04 2007 [GW][I]:         name      : builtin
>
> Wed Nov 14 07:54:04 2007 [GW][I]:         executable: gw_sched
>
> Wed Nov 14 07:54:04 2007 [GW][I]:         argument  :
>
> Wed Nov 14 07:54:04 2007 [GW][I]:
> ---------------------------------------------------
>
> Wed Nov 14 07:54:04 2007 [GW][I]:             sched.conf built-in policies
>
> Wed Nov 14 07:54:04 2007 [GW][I]:
> ---------------------------------------------------
>
> Wed Nov 14 07:54:04 2007 [GW][I]:   Scheduler configuration
> attributes
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     DISABLE                  : no
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     DISPATCH_CHUNK           : 15
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     MAX_RUNNING_USER         : 30
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     MAX_RUNNING_RESOURCE     : 10
>
> Wed Nov 14 07:54:04 2007 [GW][I]:   Job Fixed Priority Policy
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     FP_WEIGHT                : 1.00
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     Fixed Priority Values (users)
>
> Wed Nov 14 07:54:04 2007 [GW][I]:       DEFAULT                : 0
>
> Wed Nov 14 07:54:04 2007 [GW][I]:   Job Share Policy
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     SH_WEIGHT (share)        : 1.00
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     SH_WINDOW_SIZE           : 1.00
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     SH_WINDOW_DEPTH          : 5
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     User Shares
>
> Wed Nov 14 07:54:04 2007 [GW][I]:       DEFAULT                : 5
>
> Wed Nov 14 07:54:04 2007 [GW][I]:   Job Waiting time Policy
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     WT_WEIGHT                : 0.00
>
> Wed Nov 14 07:54:04 2007 [GW][I]:   Job Deadline Policy
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     DL_WEIGHT (deadline)     : 1.00
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     DL_HALF                  : 0
>
> Wed Nov 14 07:54:04 2007 [GW][I]:   Resource Fixed Priority Policy
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     RP_WEIGHT                : 1.00
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     Fixed Priority Values (information
> managers)
>
> Wed Nov 14 07:54:04 2007 [GW][I]:       DEFAULT                : 1
>
> Wed Nov 14 07:54:04 2007 [GW][I]:   Resource Failure Rate Policy
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     RA_WEIGHT                : 1.00
>
> Wed Nov 14 07:54:04 2007 [GW][I]:   Resource Failure Rank Policy
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     FR_MAX_BANNED            : 3600
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     FR_BANNED_C              : 650.00
>
> Wed Nov 14 07:54:04 2007 [GW][I]:   Resource Usage Policy
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     UG_WEIGHT                : 1.00
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     UG_HISTORY_WINDOW        : 3.00
>
> Wed Nov 14 07:54:04 2007 [GW][I]:     UG_HISTORY_RATIO         : 0.25
>
> Wed Nov 14 07:54:04 2007 [GW][I]:
> ---------------------------------------------------
>
> Wed Nov 14 07:54:04 2007 [DM][I]: Job pool initialized.
>
> Wed Nov 14 07:54:04 2007 [DM][I]: Array pool initialized.
>
> Wed Nov 14 07:54:04 2007 [IM][I]: Host pool initialized.
>
> Wed Nov 14 07:54:04 2007 [UM][I]: User pool initiated.
>
> Wed Nov 14 07:54:04 2007 [GW][I]: Loading Information Manager MADs.
>
> Wed Nov 14 07:54:05 2007 [IM][I]:       MAD mds4 loaded (exec:
> gw_im_mad_mds4_thr, arg: -l host.list).
>
> Wed Nov 14 07:54:05 2007 [GW][I]: Loading the scheduler.
>
> Wed Nov 14 07:54:05 2007 [DM][I]:       Scheduler builtin loaded (exec:
> gw_sched, arg: ).
>
> Wed Nov 14 07:54:05 2007 [GW][I]: Recovering GW state.
>
> Wed Nov 14 07:54:05 2007 [DM][I]: Dispatch Manager started.
>
> Wed Nov 14 07:54:05 2007 [TM][I]: Transfer Manager started.
>
> Wed Nov 14 07:54:05 2007 [EM][I]: Execution Manager started.
>
> Wed Nov 14 07:54:05 2007 [IM][I]: Information Manager started.
>
> Wed Nov 14 07:54:05 2007 [UM][I]: User Manager started.
>
> Wed Nov 14 07:54:05 2007 [RM][I]: Request Manager started.
>
> Wed Nov 14 07:54:10 2007 [IM][I]: Discovering hosts.
>
> Wed Nov 14 07:54:10 2007 [IM][I]: Hosts discovered by MAD (mds4):
> grid2ag.g2dominio.com grid1ag.g1dominio.com
>
>
>
> The host.list file
>
>
>
> [scheduler at scheduler ~]$ cat /usr/local/gw/host.list
>
> grid2ag.g2dominio.com
>
> grid1ag.g1dominio.com
>
>
>
>
> [scheduler at scheduler ~]$ gwhost
>
> HID PRIO  OS              ARCH   MHZ %CPU  MEM(F/T)     DISK(F/T)
> N(U/F/T) LRMS                 HOSTNAME
>
> 0   1     NULLNULL        NULL     0    0       0/0           0/0
> 0/0/1 Fork                 grid2ag.g2dominio.com
>
> 1   1     NULLNULL        NULL     0    0       0/0           0/0
> 0/0/1 Fork                 grid1ag.g1dominio.com
>
> Seems like the command gw_im_mad_mds4_thr doesn't obtain the attributes
> of the hosts.
>
>
>
> When I try to send a simple job
>
> [scheduler at scheduler test]$ gwsubmit jt
>
> [scheduler at scheduler test]$ gwps
>
> USER         JID DM   EM   START    END      EXEC    XFER    EXIT
> NAME            HOST
>
> scheduler    0   pend ---- 07:56:10 --:--:-- 0:00:00 0:00:00 --
> jt              --
>
>
>
> Another problem is that I didn't install Globus with support for
> prewsmds. I can't find the grid-info-search or globus-mds commands. How
> can I install this characteristics without reinstall all Globus?.
>
>
>
> Thanks in advance
>
>
>
> jfontan escribió:
> >
> > Hello,
> >
> > Information mad write some temp files at $GW_LOCATION/var. As you can
> > see in log files it does not have permissions (or perhaps the
> > directory doesn't even exist.):
> >
> > --8<------
> > MONITOR error in MAD (mds2): Can't access file (cwd is /usr/local/gw/var)
> > ------>8--
> >
> > Check the permissions of that directory so scheduler user does have
> > permission to read, write and list (execute).
> >
> > The problem with gwsubmit is even stranger. That error about the proxy
> > file is written when the execution manager dies after launching it
> > (the most provably reason is that it can not find the proxy file and
> > dies). But I can see that the execution manager already started in the
> > log file. Check if the PATH is correctly set and you are not mixing
> > paths and such with another gridway installation (that could be
> > already installed with globus).
> >
> > As as side note you have TM mad configured two times:
> >
> > --8<------
> >> TM_MAD = gridftp:gw_tm_mad_ftp:
> >
> >> TM_MAD = gridftp:gw_tm_mad_ftp:
> > ------>8--
> >
> > You only need one of those lines.
> >
> > Bye
> >
> >
> >
> > On Nov 13, 2007, at 8:05 PM, Eddy Diaz wrote:
> >
> >> Hi, I'm Master student in Ingeniería de sistemas y computación at the
> >> Andes University.
> >>
> >>
> >> I'm finishing my thesis and I'm trying to do an integration in real
> >> time of Grid Technologies and Collaborative tools.
> >>
> >> I've installed Globus as a middlware, Gridway as a meta-scheduler and
> >> AccessGrid as a collaborative tool.
> >>
> >> I have several problems with the gridway installation.
> >>
> >> The machine with meta-scheduler has installed fedora core 4 and
> >> Globus 4.0.5. The Globus user and the gridway user is "scheduler". I
> >> installed gridway in a single-user mode.
> >>
> >> [scheduler at scheduler gw-5.2.3]$ ./configure --prefix=/usr/local/gw/
> >> --with-docs --with-tests
> >> [scheduler at scheduler gw-5.2.3]$ make
> >> [scheduler at scheduler gw-5.2.3]$ make install
> >>
> >> The configuration file
> >>
> >> [scheduler at scheduler ~]$ vi /usr/local/gw/etc/gwd.conf
> >>
> >> IM_MAD = mds4:gw_im_mad_mds4_thr:-s grid1ag.g1dominio.com:gridftp:ws
> >> EM_MAD = ws:gw_em_mad_ws::rsl2
> >> TM_MAD = gridftp:gw_tm_mad_ftp:
> >>
> >> IM_MAD = mds2:gw_im_mad_static:-l
> >> /home/scheduler/host.list:gridftp:prews
> >> EM_MAD = prews:gw_em_mad_prews::rsl
> >> TM_MAD = gridftp:gw_tm_mad_ftp:
> >>
> >>
> >> The host.list file
> >> grid2ag.g2dominio.com
> >>
> >> When I launch the gwd daemon
> >>
> >> [scheduler at scheduler ~]$ gwd
> >>
> >> In the log file /usr/local/gw/etc/gwd.conf appears:
> >>
> >> Tue Nov 13 10:27:45 2007 [GW][I]: Loading Information Manager MADs.
> >> Tue Nov 13 10:27:46 2007 [IM][I]: MAD mds2 loaded (exec:
> >> gw_im_mad_static, arg: -l /home/scheduler/host.list).
> >> Tue Nov 13 10:27:47 2007 [IM][I]: MAD mds4 loaded (exec:
> >> gw_im_mad_mds4_thr, arg: -s grid1ag.g1dominio.com).
> >> Tue Nov 13 10:27:47 2007 [GW][I]: Loading the scheduler.
> >> Tue Nov 13 10:27:47 2007 [DM][I]: Scheduler builtin loaded (exec:
> >> gw_sched, arg: ).
> >> Tue Nov 13 10:27:47 2007 [GW][I]: Recovering GW state.
> >> Tue Nov 13 10:27:47 2007 [DM][I]: Dispatch Manager started.
> >> Tue Nov 13 10:27:47 2007 [TM][I]: Transfer Manager started.
> >> Tue Nov 13 10:27:47 2007 [EM][I]: Execution Manager started.
> >> Tue Nov 13 10:27:47 2007 [IM][I]: Information Manager started.
> >> Tue Nov 13 10:27:47 2007 [UM][I]: User Manager started.
> >> Tue Nov 13 10:27:47 2007 [RM][I]: Request Manager started.
> >> Tue Nov 13 10:27:52 2007 [IM][I]: Discovering hosts.
> >> Tue Nov 13 10:27:52 2007 [IM][I]: Hosts discovered by MAD (mds2):
> >> grid2ag.g2dominio.com
> >> Tue Nov 13 10:27:57 2007 [IM][E]: MONITOR error in MAD (mds2): Can't
> >> access file (cwd is /usr/local/gw/var)
> >> Tue Nov 13 10:28:03 2007 [IM][E]: DISCOVER error in MAD (mds4):
> >> FAILURE Error while obtaining hosts names:
> >>
> >> The list servers
> >>
> >> [scheduler at scheduler ~]$ gwhost
> >> HID PRIO OS ARCH MHZ %CPU MEM(F/T) DISK(F/T) N(U/F/T) LRMS HOSTNAME
> >> 0 1 0 0 0/0 0/0 0/0/0 grid2ag.g2dominio.com
> >>
> >> When I try to send a simple job, appears this error.
> >>
> >> [scheduler at scheduler test]$grid-proxy-init
> >> [scheduler at scheduler test]$ gwsubmit jt
> >> FAILED: failed could not register user (check proxy)
> >>
> >> I did all the "Verifying Globus Installation" in the tutorial
> >> http://www.gridway.org/documentation/files/Tutorial/2-Installation_and_Basic_Configuration-v1.0.pdf
> >> with both servers grid1ag.g1dominio.com and grid2ag.g2dominio.com and
> >> works fine.
> >>
> >> [scheduler at scheduler ~]$ globus-job-run grid2ag.g2dominio.com
> >> /bin/uname -a
> >>
> >> Linux grid2ag.g2dominio.com 2.6.11-1.1369_FC4 #1 Thu Jun 2 22:55:56
> >> EDT 2005 i686 i686 i386 GNU/Linux
> >>
> >>
> >> [scheduler at scheduler ~]$ globusrun-ws -submit -F
> >> grid1ag.g1dominio.com -s -c /bin/uname -a
> >>
> >> Delegating user credentials...Done.
> >>
> >> Submitting job...Done.
> >>
> >> Job ID: uuid:ae70092e-9202-11dc-be93-000c29f6264f
> >>
> >> Termination time: 11/14/2007 16:08 GMT
> >>
> >> Current job state: Active
> >>
> >> Current job state: CleanUp-Hold
> >>
> >> Linux grid1ag.g1dominio.com 2.6.11-1.1369_FC4 #1 Thu Jun 2 22:55:56
> >> EDT 2005 i686 i686 i386 GNU/Linux
> >>
> >> Current job state: CleanUp
> >>
> >> Current job state: Done
> >>
> >> Destroying job...Done.
> >>
> >> Cleaning up any delegated credentials...Done.
> >>
> >> Could anyone help me?
> >>
> >> Thanks in advance
> >>
> >
> >
>
>




More information about the gridway-user mailing list