Changes between Version 31 and Version 32 of GENIRacksHome/ExogeniRacks/AggregateSurvey


Ignore:
Timestamp:
04/24/12 15:48:10 (12 years ago)
Author:
lnevers@bbn.com
Comment:

--

Legend:

Unmodified
Added
Removed
Modified

  • GENIRacksHome/ExogeniRacks/AggregateSurvey

    v31 v32  
    1515
    1616== AM API Acceptance Tests ==
    17 The AM API acceptance test reported 4 failure and 2 errors this is different than the previous survey which showed 6 failures.
     17The AM API acceptance test reported 4 failure and 2 errors. This is different than the previous survey which showed 6 failures, need GPO Developement team analysis for the AM API results.
    1818
    1919{{{
    … …  
    105105== Omni Commands ==
    106106
    107 The following section capture a two node experiment (one at RENCI one at BBN).  Several tries were required to get the two VM (one per rack) working.
    108 
    109 There were a couple of converter problems, also some information about a slice seemed to be cached. Once the experiment was set up, the nodes were not able to communicate due to an OF switch problem that required a reboot.
     107The following section capture a two node experiment (one at RENCI one at BBN).  Several tries were required to get the two VMs (one per rack) up and exchanging traffic.  In the various attempts there were various issues:
     108  * Ran into a couple of NDL converter problems
     109  * Some information about a slice seemed to be cached somewhere in the environment which caused some stale state information to interfere with the experiment. Had to create a new slice and sliver to get around this provlem.
     110  * When the experiment was eventually set up, the nodes were not able to communicate due to an !OpenFlow switch problem that required a reboot. This failure occurred twice.
    110111
    111112=== List Resources ===
    112 The out from the listresources was reviewed to write the RSpec for a two ExoGENI sites (BBN and RENCI) experiment.  At first, not sure which is the RENCI Component manager ID for the RENCI ExoGENI. This is the list of potential choices: acisrencivmsite, bbnvmsite,dukevmsite, nerscvmsite, uhoustonvmsite, rcivmsite, uncvmsite. The naming does not clearly identify the ExoGENI racks. The RENCI ExoGENI rack is ''rcivmsite'' and the BBN ExoGENI rack is ''bbnvmsite''.
     113
     114The output from the listresources was reviewed to write the RSpec for a two ExoGENI sites (BBN and RENCI) experiment. 
     115'''Minor Issue:''' At first, not sure which is the RENCI Component manager ID for the RENCI ExoGENI. This is the list of potential choices: acisrencivmsite, bbnvmsite,dukevmsite, nerscvmsite, uhoustonvmsite, rcivmsite, uncvmsite. The naming does not clearly identify the ExoGENI racks. The RENCI ExoGENI rack is ''rcivmsite'' and the BBN ExoGENI rack is ''bbnvmsite''.
    113116
    114117
    … …  
    142145
    143146=== Create Slice ===
    144 Command output:
     147No issues of feedback here. Command output:
    145148
    146149{{{
    … …  
    151154INFO:omni: ------------------------------------------------------------
    152155INFO:omni: Completed createslice:
    153 
    154156  Options as run:
    155157                aggregate: https://geni.renci.org:11443/orca/xmlrpc
    156158                framework: emulab
    157159                native: True
    158 
    159160  Args: createslice lnevers6orca
    160 
    161161  Result Summary: Created slice with Name lnevers6orca, URN urn:publicid:IDN+emulab.net+slice+lnevers6orca, Expiration 2012-04-12 00:25:14+00:00
    162  
    163 INFO:omni: ============================================================
    164 }}}
    165 
     162INFO:omni: ============================================================
     163}}}
     164
     165
     166'''Note, the sections below are listed in reverse-chronological order with the final attempt listed first.'''
    166167
    167168=== Create Sliver Try-4 ===
    168 It seems that re-using the slice name cause the nodes in the same rack failure described in Try-3, so deleted sliver and slice and create a new slice name (lnevers9orca) for the 2 site sliver:
     169
     170'''Issue:''' It seems that re-using the slice name caused for the nodes in the same rack to fail as described in Try-3, so deleted sliver and slice and create a new slice name (lnevers9orca) for the 2 site sliver:
    169171{{{
    170172lnevers@sendaria:~/gcf-1.6.1$ ./src/omni.py -a  https://geni.renci.org:11443/orca/xmlrpc  -n createsliver lnevers9orca ./orca-2site.rspec
    … …  
    247249INFO:omni: ------------------------------------------------------------
    248250INFO:omni: Completed createsliver:
    249 
    250251  Options as run:
    251252                aggregate: https://geni.renci.org:11443/orca/xmlrpc
    252253                framework: emulab
    253254                native: True
    254 
    255255  Args: createsliver lnevers9orca ./orca-2site.rspec
    256 
    257256  Result Summary: Slice urn:publicid:IDN+emulab.net+slice+lnevers9orca expires within 1 day(s) on 2012-04-12 02:11:47 UTC
    258257Reserved resources on https://geni.renci.org:11443/orca/xmlrpc. 
    259258INFO:omni: ============================================================
    260 
    261259}}}
    262260
    263261=== List Resources for Sliver Try-4 ===
     262NO issues for this command:
    264263{{{
    265264lnevers@sendaria:~/gcf-1.6.1$ ./src/omni.py -a  https://geni.renci.org:11443/orca/xmlrpc  -n listresources  lnevers9orca INFO:omni:Loading config file omni_config
    … …  
    369368INFO:omni: ------------------------------------------------------------
    370369INFO:omni: Completed listresources:
    371 
    372370  Options as run:
    373371                aggregate: https://geni.renci.org:11443/orca/xmlrpc
    374372                framework: emulab
    375373                native: True
    376 
    377374  Args: listresources lnevers9orca
    378 
    379375  Result Summary: Retrieved resources for slice lnevers9orca from 1 aggregates.
    380376Wrote rspecs from 1 aggregates.
    … …  
    384380
    385381
    386 
    387382=== Login to the nodes Try-4 ===
     383
    388384The following nodes were assigned to the sliver in try-4:
    389  * Node geni2 @ rcivmsite:       control addr: "152.54.14.75" data plane addr: "172.16.1.1"
     385 * Node geni2 @ rcivmsite:       control addr: "152.54.14.75"  data plane addr: "172.16.1.1"
    390386 * Node geni1 @  bbnvmsite:      control addr:"192.1.242.15"   dataplane addr: "172.16.1.2"
    391387
    392 At first ran into a problem where the nodes did not exchange traffic which was resolved by rebooting the OpenFlow Switches at each site.
     388'''ISSUE:''' first ran into a problem where the nodes did not exchange traffic which was resolved by rebooting the OpenFlow Switches at each site.
    393389
    394390Logged into each RENCI node geni2:
    … …  
    500496=== Create Sliver Try-3 ===
    501497
    502 After another converter issue was resolved which was potentially a memory leak related to redeploying it into tomcat multiple times (not during runtime).  Issue was resolved, deleted sliver and tried to re-create the sliver, only to find that the VMs were both at BBN site again:
     498After another converter issue was resolved which was potentially a memory leak related to redeploying it into tomcat multiple times (not during runtime), deleted sliver and tried to re-create the sliver. '''Issue:'' Found out that the VMs were both at BBN site again, rather than as requested (1 at BBN and 1 at RENCI):
    503499
    504500{{{
    … …  
    601597=== Create Sliver Try-2 ===
    602598
    603 After Ilia removed converter hack (see try-1), deleted sliver and tried to re-create the sliver but ran into the following NDL converter error:
     599After Ilia removed converter hack (see try-1), deleted sliver and tried to re-create the sliver. '''Issue:''' Ran into the following NDL converter error:
    604600
    605601{{{
    … …  
    631627}}}
    632628=== Create Sliver Try-1 ==
    633 
    634 Create a sliver with one Compute Resource at BBN and one Compute Resource at RENCI. But the nodes assigned were both on the bbnvmsite.  Ilia checking converter and found old hack, which he removed..
     629Created a sliver with one Compute Resource at BBN and one Compute Resource at RENCI.
     630'''Issue:'''  Each of the 2 nodes assigned were both on the bbnvmsite.  Ilia checking converter and found old hack, which he removed..
    635631
    636632{{{
    … …  
    722718
    723719=== Sliver Status Try-1 ===
     720No issues found.
     721
    724722{{{
    725723lnevers@sendaria:~/gcf-1.6.1$ ./src/omni.py -a  https://geni.renci.org:11443/orca/xmlrpc  -n sliverstatus lnevers6orca
    … …  
    747745
    748746=== List Resources for Slice Try-1 ===
    749 To determine which node were in the slice, but found both VM at BBN site:
     747
     748'''Issue:''' To determine which node were in the slice, but found both VM at BBN site:
    750749
    751750{{{
    … …  
    806805
    807806}}}
    808 
    809807
    810808