rkhamis / testmigration Goto Github PK

create a jumpscript which exports all relevant OSIS objects to YAML nicely structured in a directory layout

cbroker & vfw

use subdirectory structure

space

• machine
• ...
  cpunode
• ...???

...

be logical info only once

this will be used as backup but also to see how data changed over time
will be put in coderepo

jumpscale extension in libs repo

use postgresql db as source of data

create methods to find events, get event, delete events based on certain criteria

(JSBOX)root@admin1:~# jsgrid disablenode -nid 10
sandbox:/opt/jsbox
ERROR
list indices must be integers, not str
Traceback (most recent call last):
~ File "/opt/jsbox/tools/jsgrid", line 279, in
_deleteNode(node)
~ File "/opt/jsbox/tools/jsgrid", line 179, in _deleteNode
nid = node['id']
~ TypeError: list indices must be integers, not str

type/level: UNKNOWN/critical
TypeError: list indices must be integers, not str

something to do with .

which goes over all installed jpackages, checks if buildnr changed, if so do an install (so really acts like update)

In [9]: cl = j.core.osis.getClient(user='root')

In [10]: accounts = acclient.simpleSearch({'name': '[email protected]'})

In [11]: acclient.get(45)
Out[11]:
{
"_ckey": "",
"_meta": [
"osismodel",
"cloudbroker",
"account",
1
],
"acl": [
{
"guid": "",
"right": "CXDRAU",
"type": "U",
"userGroupId": "[email protected]"
}
],
"guid": 45,
"id": 45,
"name": "[email protected]"
}

In [12]: accounts = acclient.simpleSearch({'name': '[email protected]'})

In [13]: acclient.simpleSearch({'name': '[email protected]'})
Out[13]: []

i've added in jsgrid possibility to remove jobs.ecos
doesn't seem to work, need this urgently

make new graphs (these are fake)

• show nr of vmachines
• show nr of cpu nodes
• show nr of accounts/spaces
• status of health but now not layout nice

as we do updates regularly, can we make sure the portal and the proxy are avoiding caching, this can be done by finetuning the caching headers

apache example:

<filesMatch ".(css|js|pdf)$">
Header set Cache-Control "max-age=0, no-cache, no-store, must-revalidate"

When specifying a queue in the jumpscripst itself it does not execute on the correct queue.

i've added in jsgrid possibility to remove jobs.ecos
doesn't seem to work, need this urgently

body

• document how to use gridportal to identify cpu resource issues
• create new page & macro to show CPU usage per cpu and fw node
  ** show in 1 page, this allows us to see immediate which cpu/fw node is overloaded
  ** cpu usage means context switches (define danger zone) & cpu usage
  ** the 1 page lists all cpu/fw nodes & their cpu status
• there should be a page per compute node which shows cpu usage per VM
  ** is not the main page of the compute node, its a 2nd one only called when user wants to troubleshoot resource utilization
  ** lists all vm's & their used mem,cpu,bandwidth? ... whatever relevant to see resource used
• info comes from graphite or our aggregator or osis (we have multiple locations with this info)
• check that our monitoring is working properly
  ...

Like this one :
http://gm1.lenoir2.vscalers.com:81/grid/eco?id=301_2_1768

Users could be created cross grid regardeless they should be changeable on both grids.

Eg. create user Jo on lenoir2 key == 400_jo
change him om york1 key == 301_jo things dont work

http://support.mothership1.com/kb/faq.php?id=40
we need to mention we are only active in few DC's

Run on cb1.york1.vscalers.com:

{code}
In [2]: cl = j.core.osis.getClient(user='root')

In [6]: vmcl = j.core.osis.getClientForCategory(cl, 'cloudbroker', 'vmachine')

In [7]: vmcl.get(357)
SPECPATH:/opt/jsbox_data/var/code/osis/cloudbroker/model.spec
getOsisModelClass: cloudbroker_vmachine
Out[7]: ---------------------------------------------------------------------------
UnicodeEncodeError Traceback (most recent call last)
/root/cloudbroker_shell in ()
----> 1 vmcl.get(357)

/opt/jsbox/lib/IPython/core/displayhook.pyc in call(self, result)
236 self.start_displayhook()
237 self.write_output_prompt()
--> 238 format_dict = self.compute_format_data(result)
239 self.write_format_data(format_dict)
240 self.update_user_ns(result)

/opt/jsbox/lib/IPython/core/displayhook.pyc in compute_format_data(self, result)
148 MIME type representation of the object.
149 """
--> 150 return self.shell.display_formatter.format(result)
151
152 def write_format_data(self, format_dict):

/opt/jsbox/lib/IPython/core/formatters.pyc in format(self, obj, include, exclude)
124 continue
125 try:
--> 126 data = formatter(obj)
127 except:
128 # FIXME: log the exception

/opt/jsbox/lib/IPython/core/formatters.pyc in call(self, obj)
445 type_pprinters=self.type_printers,
446 deferred_pprinters=self.deferred_printers)
--> 447 printer.pretty(obj)
448 printer.flush()
449 return stream.getvalue()

/opt/jsbox/lib/IPython/lib/pretty.pyc in pretty(self, obj)
358 if callable(meth):
359 return meth(obj, self, cycle)
--> 360 return _default_pprint(obj, self, cycle)
361 finally:
362 self.end_group()

/opt/jsbox/lib/IPython/lib/pretty.pyc in _default_pprint(obj, p, cycle)
478 if getattr(klass, 'repr', None) not in _baseclass_reprs:
479 # A user-provided repr.
--> 480 p.text(repr(obj))
481 return
482 p.begin_group(1, '<')

UnicodeEncodeError: 'ascii' codec can't encode character u'\xe9' in position 423: ordinal not in range(128)

In [8]:
{code}

eg. add _reloadjumpscript to args of jumpscript which will bypass cache

just that: unclickable links

In the grid portal for very vm, when it has ever been created, a nic entry is generated that lives forever in OSIS.

nice feature, would be great to see linenrs on code

grid should be with search per col if above macro is used
and should be datatables
please fix this macro

/opt/jsbox/apps/redis/redis-check-aof --fix /opt/jsbox_data/var/redis/redism/db/appendonly.aof

nice feature, would be great to see linenrs on code

something to do with .

5 jumpscale:processmanager 14:58:46: exec:netstat -ntulp | grep ':7768 '
5 jumpscale:processmanager 14:58:46: exec:env COLUMNS=300 ps -ef
5 jumpscale:processmanager 14:58:46: exec:env COLUMNS=300 ps -ef
5 jumpscale:processmanager 14:58:46: exec:netstat -ntulp | grep ':7766 '
5 jumpscale:processmanager 14:58:46: exec:env COLUMNS=300 ps -ef
5 jumpscale:processmanager 14:58:46: exec:env COLUMNS=300 ps -ef
5 jumpscale:processmanager 14:58:46: exec:netstat -ntulp | grep ':7767 '
5 jumpscale:processmanager 14:58:46: exec:env COLUMNS=300 ps -ef
5 jumpscale:processmanager 14:58:46: exec:env COLUMNS=300 ps -ef
5 jumpscale:processmanager 14:58:46: exec:netstat -ntulp | grep ':7769 '
5 jumpscale:processmanager 14:58:46: exec:env COLUMNS=300 ps -ef
5 jumpscale:processmanager 14:58:46: exec:env COLUMNS=300 ps -ef

why not use a delete query?
please do a query which counts before the delete & after, show how many where deleted

http://gm1.lenoir2.vscalers.com:81/grid/jobs?state=error

see
file:///opt/jumpscale/apps/gridportal/base/Grid/.macros/wiki/alert/1_alert.py
get info out of redis

see file:///opt/jumpscale/apps/watchdogmanager/populate_redis_test.py
to see how redis is used

needs to work together with
file:///opt/jumpscale/apps/gridportal/base/Grid/.macros/wiki/alerts/1_alerts.py
make sure alerts is working as well

to run locally do first the populate_redis_test
then run
file:///opt/jumpscale/apps/watchdogmanager/watchdogManager.py
to have alerts

Check what we can do with old pids

Harden the macro to print error instead of making the whole portal crash