Hello
we have this strange behaviour…
some of our esx have access to the same datastores… is a cluster… normal i think
the issue is that some esx report a datastore as critical while the others do not.
this is the status of a critical :
COMP-02-VMFS-CON-T1=1540364.0MB;1493977.6;1525429.76;0;1572608 growth=0;;;; trend=0;;;0;65525.333333 uncommitted=0.0MB;;;; overprovisioned=1540364.0MB;;;;
this is the same datastore in another esx that is not critical:
COMP-02-VMFS-CON-T1=1433754.0MB;1493977.6;1525429.76;0;1572608 growth=0;;;; trend=13404.104882;;;0;65525.333333 uncommitted=0.0MB;;;; overprovisioned=1433754.0MB;;;;
the only difference I can see is the overprovisioned… what exactly is this?
···
Mario Garcia
ABSI NV | System Engineer
Humaniteitslaan 224 - Bld de l’Humanité - 1190 Bruxelles
mariog@absi.be / http://www.absi.be / Tel: +32(0)2 333 40 00 Fax: +32(0)2 333 40 60
Hello,
overprovisioned is important if you make the vmdk-files thin- instead of thick provisioned.
Example: lets say you have a VMFS datastore with 200GB size. You make a new VM with a total harddrive capacitiy of 300GB and thin provision.
Then the size of the vmdk files is as big as they just need to be and the datastore is overprovisioned by 100GB. If the VM some day needs more than 200GB and grows its vmdk files, then the Vm will stop working as soon as the 200Gb are full. so Check_MK has the ability to show this overprovision and warn you.
Then, if you monitor your esx hosts directly (not the vcenter), then the overprovision data is always the same as the actual occupied space. I can see this in your output too.
For correct datastore monitoring i suggest you disable datastore monitoring on your esxi host checks and enable the datastore monitoring on the vcenter instead.
not only you just have one alert for warn/crit if it happens (instead of x times where is is the number of hosts which have the same datastore mounted),
the vcenter monitoring displays the correct overprovision of a datastore.
Example:
This is what Check_MK sees on the vcenter (correct):
LUN160_data2=132940.0MB;1922532.491138;1963245.660759;0;2044672 growth=0;;;; trend=-461.556376;;;0;85194.666667 trend_hoursleft=-1;;;; uncommitted=1905813.00058MB;;;; overprovisioned=2038753.00058MB;;;;;;
This is what Check_MK sees on one of the esxi hosts which have this datastore mounted:
LUN160_data2=132940.0MB;1922532.491138;2003958.830379;0;2044672 growth=0;;;; trend=-15.200769;;;0;85194.666667 trend_hoursleft=-1;;;; uncommitted=0.0MB;;;; overprovisioned=132940.0MB;;;;
And yes, sometimes i have the same phenomenon: that different esxi hosts report different metrics on the same datastore. this is why i make it this way now.
···
2016-04-08 14:05 GMT+02:00 Mario Garcia mariog@absi.be:
Hello
we have this strange behaviour…
some of our esx have access to the same datastores… is a cluster… normal i think
the issue is that some esx report a datastore as critical while the others do not.
this is the status of a critical :
COMP-02-VMFS-CON-T1=1540364.0MB;1493977.6;1525429.76;0;1572608 growth=0;;;; trend=0;;;0;65525.333333 uncommitted=0.0MB;;;; overprovisioned=1540364.0MB;;;;
this is the same datastore in another esx that is not critical:
COMP-02-VMFS-CON-T1=1433754.0MB;1493977.6;1525429.76;0;1572608 growth=0;;;; trend=13404.104882;;;0;65525.333333 uncommitted=0.0MB;;;; overprovisioned=1433754.0MB;;;;
–
Mario Garcia
ABSI NV | System Engineer
Humaniteitslaan 224 - Bld de l’Humanité - 1190 Bruxelles
mariog@absi.be / http://www.absi.be / Tel: +32(0)2 333 40 00 Fax: +32(0)2 333 40 60
the only difference I can see is the overprovisioned… what exactly is this?
checkmk-en mailing list
checkmk-en@lists.mathias-kettner.de
http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en
This happens as every ESX host only knows the values of actively used datastores. If you have no running machine on one datastores then you see some older values on this VMFS. That’s a normal behavior.
Only way around is to grab the datastore value from vcenter.
Best regards
Andreas
···
Mario Garcia mariog@absi.be schrieb am Fr., 8. Apr. 2016, 14:07:
Hello
we have this strange behaviour…
some of our esx have access to the same datastores… is a cluster… normal i think
the issue is that some esx report a datastore as critical while the others do not.
this is the status of a critical :
COMP-02-VMFS-CON-T1=1540364.0MB;1493977.6;1525429.76;0;1572608 growth=0;;;; trend=0;;;0;65525.333333 uncommitted=0.0MB;;;; overprovisioned=1540364.0MB;;;;
this is the same datastore in another esx that is not critical:
COMP-02-VMFS-CON-T1=1433754.0MB;1493977.6;1525429.76;0;1572608 growth=0;;;; trend=13404.104882;;;0;65525.333333 uncommitted=0.0MB;;;; overprovisioned=1433754.0MB;;;;
the only difference I can see is the overprovisioned… what exactly is this?
–
Mario Garcia
ABSI NV | System Engineer
Humaniteitslaan 224 - Bld de l’Humanité - 1190 Bruxelles
mariog@absi.be / http://www.absi.be / Tel: +32(0)2 333 40 00 Fax: +32(0)2 333 40 60
checkmk-en mailing list
checkmk-en@lists.mathias-kettner.de
http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en
Hello
thanks for all these insight… very useful indeed.
is there a particular setting to only allow datastore monitoring in a vcenter or it has to be done with tags?
thanks
Mario.
···
On 9 April 2016 at 06:53, Andreas Döhler andreas.doehler@gmail.com wrote:
This happens as every ESX host only knows the values of actively used datastores. If you have no running machine on one datastores then you see some older values on this VMFS. That’s a normal behavior.
Only way around is to grab the datastore value from vcenter.
Best regards
Andreas
Mario Garcia mariog@absi.be schrieb am Fr., 8. Apr. 2016, 14:07:
Hello
we have this strange behaviour…
some of our esx have access to the same datastores… is a cluster… normal i think
the issue is that some esx report a datastore as critical while the others do not.
this is the status of a critical :
COMP-02-VMFS-CON-T1=1540364.0MB;1493977.6;1525429.76;0;1572608 growth=0;;;; trend=0;;;0;65525.333333 uncommitted=0.0MB;;;; overprovisioned=1540364.0MB;;;;
this is the same datastore in another esx that is not critical:
COMP-02-VMFS-CON-T1=1433754.0MB;1493977.6;1525429.76;0;1572608 growth=0;;;; trend=13404.104882;;;0;65525.333333 uncommitted=0.0MB;;;; overprovisioned=1433754.0MB;;;;
the only difference I can see is the overprovisioned… what exactly is this?
–
Mario Garcia
ABSI NV | System Engineer
Humaniteitslaan 224 - Bld de l’Humanité - 1190 Bruxelles
mariog@absi.be / http://www.absi.be / Tel: +32(0)2 333 40 00 Fax: +32(0)2 333 40 60
checkmk-en mailing list
checkmk-en@lists.mathias-kettner.de
http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en
–
Mario Garcia
ABSI NV | System Engineer
Humaniteitslaan 224 - Bld de l’Humanité - 1190 Bruxelles
mariog@absi.be / http://www.absi.be / Tel: +32(0)2 333 40 00 Fax: +32(0)2 333 40 60
WATO->Host & Service Parameters->Datasource Programs->Check state of VMWare ESX via vSphere
Then create a rule for the vcenter and enable the metrics you want to have monitored:
Host Systems
Virtual Machines
Datastores
Performance Counters
License Usage
Tom
P.S: of course you need a readonly user defined on top of the vcenter. 
···
2016-04-11 10:22 GMT+02:00 Mario Garcia mariog@absi.be:
Hello
thanks for all these insight… very useful indeed.
is there a particular setting to only allow datastore monitoring in a vcenter or it has to be done with tags?
thanks
Mario.
checkmk-en mailing list
checkmk-en@lists.mathias-kettner.de
http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en
On 9 April 2016 at 06:53, Andreas Döhler andreas.doehler@gmail.com wrote:
This happens as every ESX host only knows the values of actively used datastores. If you have no running machine on one datastores then you see some older values on this VMFS. That’s a normal behavior.
Only way around is to grab the datastore value from vcenter.
Best regards
Andreas
Mario Garcia mariog@absi.be schrieb am Fr., 8. Apr. 2016, 14:07:
Hello
we have this strange behaviour…
some of our esx have access to the same datastores… is a cluster… normal i think
the issue is that some esx report a datastore as critical while the others do not.
this is the status of a critical :
COMP-02-VMFS-CON-T1=1540364.0MB;1493977.6;1525429.76;0;1572608 growth=0;;;; trend=0;;;0;65525.333333 uncommitted=0.0MB;;;; overprovisioned=1540364.0MB;;;;
this is the same datastore in another esx that is not critical:
COMP-02-VMFS-CON-T1=1433754.0MB;1493977.6;1525429.76;0;1572608 growth=0;;;; trend=13404.104882;;;0;65525.333333 uncommitted=0.0MB;;;; overprovisioned=1433754.0MB;;;;
the only difference I can see is the overprovisioned… what exactly is this?
–
Mario Garcia
ABSI NV | System Engineer
Humaniteitslaan 224 - Bld de l’Humanité - 1190 Bruxelles
mariog@absi.be / http://www.absi.be / Tel: +32(0)2 333 40 00 Fax: +32(0)2 333 40 60
checkmk-en mailing list
checkmk-en@lists.mathias-kettner.de
http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en
–
Mario Garcia
ABSI NV | System Engineer
Humaniteitslaan 224 - Bld de l’Humanité - 1190 Bruxelles
mariog@absi.be / http://www.absi.be / Tel: +32(0)2 333 40 00 Fax: +32(0)2 333 40 60
yes I had missed the last part… and had no read only user…
thanks alot
···
On 11 April 2016 at 10:52, Thomas Wittmann tom.teel@gmail.com wrote:
WATO->Host & Service Parameters->Datasource Programs->Check state of VMWare ESX via vSphere
Then create a rule for the vcenter and enable the metrics you want to have monitored:
Host Systems
Virtual Machines
Datastores
Performance Counters
License Usage
Tom
P.S: of course you need a readonly user defined on top of the vcenter.
–
2016-04-11 10:22 GMT+02:00 Mario Garcia mariog@absi.be:
Hello
thanks for all these insight… very useful indeed.
is there a particular setting to only allow datastore monitoring in a vcenter or it has to be done with tags?
thanks
Mario.
checkmk-en mailing list
checkmk-en@lists.mathias-kettner.de
http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en
On 9 April 2016 at 06:53, Andreas Döhler andreas.doehler@gmail.com wrote:
This happens as every ESX host only knows the values of actively used datastores. If you have no running machine on one datastores then you see some older values on this VMFS. That’s a normal behavior.
Only way around is to grab the datastore value from vcenter.
Best regards
Andreas
Mario Garcia mariog@absi.be schrieb am Fr., 8. Apr. 2016, 14:07:
Hello
we have this strange behaviour…
some of our esx have access to the same datastores… is a cluster… normal i think
the issue is that some esx report a datastore as critical while the others do not.
this is the status of a critical :
COMP-02-VMFS-CON-T1=1540364.0MB;1493977.6;1525429.76;0;1572608 growth=0;;;; trend=0;;;0;65525.333333 uncommitted=0.0MB;;;; overprovisioned=1540364.0MB;;;;
this is the same datastore in another esx that is not critical:
COMP-02-VMFS-CON-T1=1433754.0MB;1493977.6;1525429.76;0;1572608 growth=0;;;; trend=13404.104882;;;0;65525.333333 uncommitted=0.0MB;;;; overprovisioned=1433754.0MB;;;;
the only difference I can see is the overprovisioned… what exactly is this?
–
Mario Garcia
ABSI NV | System Engineer
Humaniteitslaan 224 - Bld de l’Humanité - 1190 Bruxelles
mariog@absi.be / http://www.absi.be / Tel: +32(0)2 333 40 00 Fax: +32(0)2 333 40 60
checkmk-en mailing list
checkmk-en@lists.mathias-kettner.de
http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en
–
Mario Garcia
ABSI NV | System Engineer
Humaniteitslaan 224 - Bld de l’Humanité - 1190 Bruxelles
mariog@absi.be / http://www.absi.be / Tel: +32(0)2 333 40 00 Fax: +32(0)2 333 40 60
Mario Garcia
ABSI NV | System Engineer
Humaniteitslaan 224 - Bld de l’Humanité - 1190 Bruxelles
mariog@absi.be / http://www.absi.be / Tel: +32(0)2 333 40 00 Fax: +32(0)2 333 40 60