Discussion:
number of cores
Ugo Vasi
2018-11-19 10:29:55 UTC
Permalink
Hi all,
in the dashboard of an ACS installation vesion 4.11.1.0 (Ubuntu 16.04
with KVM hypervisor), the new entry "# of CPU Cores" appears.
Is it possible to over-provision like for MHz or storage?

Thanks
--
*Ugo Vasi* / System Administrator
***@procne.it <mailto:***@procne.it>




*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>


Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da parte
di qualsiasi soggetto diverso dal destinatario è proibita sia ai sensi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete ricevuto
questo messaggio per errore, vi preghiamo di distruggerlo e di informare
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
***@procne.it <mailto:***@procne.it>.
Andrija Panic
2018-11-19 11:56:39 UTC
Permalink
Hi Ugo,

Why would you want to do this, just curious ?

I believe it's not possible, but anyway (at least with KVM, probably same
for other hypervisors) it doesn't even makes sense/use, since when
deploying a VM, ACS query host free/unused number of MHz (GHz), so it's not
even relevant for ACS - number of cores in not relevant in ACS calculations
during VM deployment.


Cheers,
Andrija
Post by Ugo Vasi
Hi all,
in the dashboard of an ACS installation vesion 4.11.1.0 (Ubuntu 16.04
with KVM hypervisor), the new entry "# of CPU Cores" appears.
Is it possible to over-provision like for MHz or storage?
Thanks
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da parte
di qualsiasi soggetto diverso dal destinatario Ú proibita sia ai sensi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete ricevuto
questo messaggio per errore, vi preghiamo di distruggerlo e di informare
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
Ugo Vasi
2018-11-19 13:11:44 UTC
Permalink
Hi Andrija,
not having noticed this new voice before I wondered if it is limiting
the fact of reaching or exceeding the number of physical cores.

What is the purpose of this dashboard pane?
Post by Andrija Panic
Hi Ugo,
Why would you want to do this, just curious ?
I believe it's not possible, but anyway (at least with KVM, probably same
for other hypervisors) it doesn't even makes sense/use, since when
deploying a VM, ACS query host free/unused number of MHz (GHz), so it's not
even relevant for ACS - number of cores in not relevant in ACS calculations
during VM deployment.
Cheers,
Andrija
Post by Ugo Vasi
Hi all,
in the dashboard of an ACS installation vesion 4.11.1.0 (Ubuntu 16.04
with KVM hypervisor), the new entry "# of CPU Cores" appears.
Is it possible to over-provision like for MHz or storage?
Thanks
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da parte
di qualsiasi soggetto diverso dal destinatario è proibita sia ai sensi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete ricevuto
questo messaggio per errore, vi preghiamo di distruggerlo e di informare
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
--
*Ugo Vasi* / System Administrator
***@procne.it <mailto:***@procne.it>




*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>


Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da parte
di qualsiasi soggetto diverso dal destinatario è proibita sia ai sensi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete ricevuto
questo messaggio per errore, vi preghiamo di distruggerlo e di informare
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
***@procne.it <mailto:***@procne.it>.
Andrija Panic
2018-11-19 14:34:41 UTC
Permalink
Unless someone gives you better answer, I guess it's for fun - to have more
detailed numbers in dashboard (may be it's related to other hypervisor
types, just assuming... or not...)

Cheers
Post by Ugo Vasi
Hi Andrija,
not having noticed this new voice before I wondered if it is limiting
the fact of reaching or exceeding the number of physical cores.
What is the purpose of this dashboard pane?
Post by Andrija Panic
Hi Ugo,
Why would you want to do this, just curious ?
I believe it's not possible, but anyway (at least with KVM, probably same
for other hypervisors) it doesn't even makes sense/use, since when
deploying a VM, ACS query host free/unused number of MHz (GHz), so it's
not
Post by Andrija Panic
even relevant for ACS - number of cores in not relevant in ACS
calculations
Post by Andrija Panic
during VM deployment.
Cheers,
Andrija
Post by Ugo Vasi
Hi all,
in the dashboard of an ACS installation vesion 4.11.1.0 (Ubuntu 16.04
with KVM hypervisor), the new entry "# of CPU Cores" appears.
Is it possible to over-provision like for MHz or storage?
Thanks
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da parte
di qualsiasi soggetto diverso dal destinatario Ú proibita sia ai sensi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete ricevuto
questo messaggio per errore, vi preghiamo di distruggerlo e di informare
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da parte
di qualsiasi soggetto diverso dal destinatario Ú proibita sia ai sensi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete ricevuto
questo messaggio per errore, vi preghiamo di distruggerlo e di informare
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
--
Andrija Panić
Dag Sonstebo
2018-11-19 15:09:51 UTC
Permalink
Ultimately there will always be a 1-to-1 relationship between vCPU cores and physical cores - you can't make a 4vCPU VM run on an "overprovisioned 2 physical core" hypervisor (unless you consider Intel HT to be exactly this). You can only overprovision time slices of your total clock cycles. If you push the limits too far on this though you get vCPU wait times - when there is too much contention and it ends up taking too long to schedule e.g. 4 vCPUs on 4 physical cores.

IMO having the total number of cores is still useful information - same as total number of hosts and memory. Yes we know these are total numbers - but still gives you an idea of the capacity of your compute nodes.

Regards,
Dag Sonstebo
Cloud Architect
ShapeBlue


On 19/11/2018, 14:43, "Andrija Panic" <***@gmail.com> wrote:

Unless someone gives you better answer, I guess it's for fun - to have more
detailed numbers in dashboard (may be it's related to other hypervisor
types, just assuming... or not...)

Cheers


***@shapeblue.com 
www.shapeblue.com
Amadeus House, Floral Street, London WC2E 9DPUK
@shapeblue
Post by Ugo Vasi
Hi Andrija,
not having noticed this new voice before I wondered if it is limiting
the fact of reaching or exceeding the number of physical cores.
What is the purpose of this dashboard pane?
Post by Andrija Panic
Hi Ugo,
Why would you want to do this, just curious ?
I believe it's not possible, but anyway (at least with KVM, probably same
for other hypervisors) it doesn't even makes sense/use, since when
deploying a VM, ACS query host free/unused number of MHz (GHz), so it's
not
Post by Andrija Panic
even relevant for ACS - number of cores in not relevant in ACS
calculations
Post by Andrija Panic
during VM deployment.
Cheers,
Andrija
Post by Ugo Vasi
Hi all,
in the dashboard of an ACS installation vesion 4.11.1.0 (Ubuntu 16.04
with KVM hypervisor), the new entry "# of CPU Cores" appears.
Is it possible to over-provision like for MHz or storage?
Thanks
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da parte
di qualsiasi soggetto diverso dal destinatario è proibita sia ai sensi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete ricevuto
questo messaggio per errore, vi preghiamo di distruggerlo e di informare
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da parte
di qualsiasi soggetto diverso dal destinatario è proibita sia ai sensi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete ricevuto
questo messaggio per errore, vi preghiamo di distruggerlo e di informare
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
--

Andri
Andrija Panic
2018-11-19 15:56:07 UTC
Permalink
Thx Dag for confirmation.

Speaking about CPU cores - I do have one "problem" which not sure is
possible to solve in meaningful way

2 x 8 core Intel Xeon E5 2.6GHz nominally (with HT enabled), with burst to
3.4 GHz ( but ! only a few cores can burst to that 3.4 - not all at the
same time !) - and CloudStack reports my host as 32 x 3.4GHz (instead of
32 x 2.6 GHz)
(not sure from where does ACS extract this 3.4GHz info... - effectively my
host CPUs are overprovisioned here by 3.4 / 2.6 = 1.30 (30%
overprovisioning)

One note, we do have CPU governor set to high performance etc...

Any opinion ?

Cheers
Post by Dag Sonstebo
Ultimately there will always be a 1-to-1 relationship between vCPU cores
and physical cores - you can't make a 4vCPU VM run on an "overprovisioned 2
physical core" hypervisor (unless you consider Intel HT to be exactly
this). You can only overprovision time slices of your total clock cycles.
If you push the limits too far on this though you get vCPU wait times -
when there is too much contention and it ends up taking too long to
schedule e.g. 4 vCPUs on 4 physical cores.
IMO having the total number of cores is still useful information - same as
total number of hosts and memory. Yes we know these are total numbers - but
still gives you an idea of the capacity of your compute nodes.
Regards,
Dag Sonstebo
Cloud Architect
ShapeBlue
Unless someone gives you better answer, I guess it's for fun - to have more
detailed numbers in dashboard (may be it's related to other hypervisor
types, just assuming... or not...)
Cheers
www.shapeblue.com
Amadeus House, Floral Street, London WC2E 9DPUK
@shapeblue
Post by Ugo Vasi
Hi Andrija,
not having noticed this new voice before I wondered if it is limiting
the fact of reaching or exceeding the number of physical cores.
What is the purpose of this dashboard pane?
Post by Andrija Panic
Hi Ugo,
Why would you want to do this, just curious ?
I believe it's not possible, but anyway (at least with KVM,
probably same
Post by Ugo Vasi
Post by Andrija Panic
for other hypervisors) it doesn't even makes sense/use, since when
deploying a VM, ACS query host free/unused number of MHz (GHz), so
it's
Post by Ugo Vasi
not
Post by Andrija Panic
even relevant for ACS - number of cores in not relevant in ACS
calculations
Post by Andrija Panic
during VM deployment.
Cheers,
Andrija
Post by Ugo Vasi
Hi all,
in the dashboard of an ACS installation vesion 4.11.1.0 (Ubuntu
16.04
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
with KVM hypervisor), the new entry "# of CPU Cores" appears.
Is it possible to over-provision like for MHz or storage?
Thanks
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i
relativi
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso
da parte
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
di qualsiasi soggetto diverso dal destinatario Ú proibita sia ai
sensi
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n.
196/2003
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
"Codice in materia di protezione dei dati personali". Se avete
ricevuto
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
questo messaggio per errore, vi preghiamo di distruggerlo e di
informare
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da
parte
Post by Ugo Vasi
di qualsiasi soggetto diverso dal destinatario Ú proibita sia ai
sensi
Post by Ugo Vasi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete
ricevuto
Post by Ugo Vasi
questo messaggio per errore, vi preghiamo di distruggerlo e di
informare
Post by Ugo Vasi
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
--
Andrija Panić
--
Andrija Panić
Yiping Zhang
2018-11-19 23:47:57 UTC
Permalink
Eric:

What's your value for global setting cpu.overprovisioning.factor?

I have this value set to 3.0. Right now, one of my servers with 32 cores @ 2.0 GHz (with HT enabled), I can allocate a total of 79 vCPU and 139 GHz to 26 VM instances. That's over 200% over provisioning!

Yiping

On 11/19/18, 6:43 AM, "Andrija Panic" <***@gmail.com> wrote:

Unless someone gives you better answer, I guess it's for fun - to have more
detailed numbers in dashboard (may be it's related to other hypervisor
types, just assuming... or not...)

Cheers
Post by Ugo Vasi
Hi Andrija,
not having noticed this new voice before I wondered if it is limiting
the fact of reaching or exceeding the number of physical cores.
What is the purpose of this dashboard pane?
Post by Andrija Panic
Hi Ugo,
Why would you want to do this, just curious ?
I believe it's not possible, but anyway (at least with KVM, probably same
for other hypervisors) it doesn't even makes sense/use, since when
deploying a VM, ACS query host free/unused number of MHz (GHz), so it's
not
Post by Andrija Panic
even relevant for ACS - number of cores in not relevant in ACS
calculations
Post by Andrija Panic
during VM deployment.
Cheers,
Andrija
Post by Ugo Vasi
Hi all,
in the dashboard of an ACS installation vesion 4.11.1.0 (Ubuntu 16.04
with KVM hypervisor), the new entry "# of CPU Cores" appears.
Is it possible to over-provision like for MHz or storage?
Thanks
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da parte
di qualsiasi soggetto diverso dal destinatario è proibita sia ai sensi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete ricevuto
questo messaggio per errore, vi preghiamo di distruggerlo e di informare
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da parte
di qualsiasi soggetto diverso dal destinatario è proibita sia ai sensi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete ricevuto
questo messaggio per errore, vi preghiamo di distruggerlo e di informare
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
--

Andrija Panić
Andrija Panic
2018-11-19 23:57:57 UTC
Permalink
FYI, tested again (CPU overprovisioning factor=1 on both Cluster and Global
level - so NO overprovisioning)

deployed VM unti it failed (could not deploy any more):

At moment of failure:
cpu used 79% (90% is disabled threshold, but VMs is big one, so would cause
91-92% usage)
number of cpu cores: 143%

So in my tests, it has nothing to do with cpu core limit (nor
overprovisioning), since I'm way above it and overprovisioning is disabled.
Post by Yiping Zhang
What's your value for global setting cpu.overprovisioning.factor?
2.0 GHz (with HT enabled), I can allocate a total of 79 vCPU and 139 GHz to
26 VM instances. That's over 200% over provisioning!
Yiping
Unless someone gives you better answer, I guess it's for fun - to have more
detailed numbers in dashboard (may be it's related to other hypervisor
types, just assuming... or not...)
Cheers
Post by Ugo Vasi
Hi Andrija,
not having noticed this new voice before I wondered if it is limiting
the fact of reaching or exceeding the number of physical cores.
What is the purpose of this dashboard pane?
Post by Andrija Panic
Hi Ugo,
Why would you want to do this, just curious ?
I believe it's not possible, but anyway (at least with KVM,
probably same
Post by Ugo Vasi
Post by Andrija Panic
for other hypervisors) it doesn't even makes sense/use, since when
deploying a VM, ACS query host free/unused number of MHz (GHz), so
it's
Post by Ugo Vasi
not
Post by Andrija Panic
even relevant for ACS - number of cores in not relevant in ACS
calculations
Post by Andrija Panic
during VM deployment.
Cheers,
Andrija
Post by Ugo Vasi
Hi all,
in the dashboard of an ACS installation vesion 4.11.1.0 (Ubuntu
16.04
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
with KVM hypervisor), the new entry "# of CPU Cores" appears.
Is it possible to over-provision like for MHz or storage?
Thanks
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i
relativi
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso
da parte
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
di qualsiasi soggetto diverso dal destinatario Ú proibita sia ai
sensi
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n.
196/2003
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
"Codice in materia di protezione dei dati personali". Se avete
ricevuto
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
questo messaggio per errore, vi preghiamo di distruggerlo e di
informare
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da
parte
Post by Ugo Vasi
di qualsiasi soggetto diverso dal destinatario Ú proibita sia ai
sensi
Post by Ugo Vasi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete
ricevuto
Post by Ugo Vasi
questo messaggio per errore, vi preghiamo di distruggerlo e di
informare
Post by Ugo Vasi
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
--
Andrija Panić
--
Andrija Panić
Eric Lee Green
2018-11-19 23:59:12 UTC
Permalink
Post by Yiping Zhang
What's your value for global setting cpu.overprovisioning.factor?
I changed it to 4.0 and restarted the management server first thing. It
started out at 1.5. At 1.5, my zone shows 41% usage with the typical
workload -- 169.60Ghz / 409.78Ghz.  I have 2x24x3.03ghz and 1x24x2.40ghz
servers for a total of 203.04Ghz actual, so even without the multiplier
I'm not over provisioning my CPU Mhz.
Post by Yiping Zhang
Unless someone gives you better answer, I guess it's for fun - to have more
detailed numbers in dashboard (may be it's related to other hypervisor
types, just assuming... or not...)
Cheers
Post by Ugo Vasi
Hi Andrija,
not having noticed this new voice before I wondered if it is limiting
the fact of reaching or exceeding the number of physical cores.
What is the purpose of this dashboard pane?
Post by Andrija Panic
Hi Ugo,
Why would you want to do this, just curious ?
I believe it's not possible, but anyway (at least with KVM, probably same
for other hypervisors) it doesn't even makes sense/use, since when
deploying a VM, ACS query host free/unused number of MHz (GHz), so it's
not
Post by Andrija Panic
even relevant for ACS - number of cores in not relevant in ACS
calculations
Post by Andrija Panic
during VM deployment.
Cheers,
Andrija
Post by Ugo Vasi
Hi all,
in the dashboard of an ACS installation vesion 4.11.1.0 (Ubuntu 16.04
with KVM hypervisor), the new entry "# of CPU Cores" appears.
Is it possible to over-provision like for MHz or storage?
Thanks
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da parte
di qualsiasi soggetto diverso dal destinatario è proibita sia ai sensi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete ricevuto
questo messaggio per errore, vi preghiamo di distruggerlo e di informare
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da parte
di qualsiasi soggetto diverso dal destinatario è proibita sia ai sensi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete ricevuto
questo messaggio per errore, vi preghiamo di distruggerlo e di informare
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
--
Andrija Panić
Andrija Panic
2018-11-20 00:03:55 UTC
Permalink
small tip - if you change cpu overprovisioning on CLUSTER level, no need to
restart management server - it's applied immediately on the fly.
Post by Yiping Zhang
Post by Yiping Zhang
What's your value for global setting cpu.overprovisioning.factor?
I have this value set to 3.0. Right now, one of my servers with 32 cores
@ 2.0 GHz (with HT enabled), I can allocate a total of 79 vCPU and 139 GHz
to 26 VM instances. That's over 200% over provisioning!
I changed it to 4.0 and restarted the management server first thing. It
started out at 1.5. At 1.5, my zone shows 41% usage with the typical
workload -- 169.60Ghz / 409.78Ghz. I have 2x24x3.03ghz and 1x24x2.40ghz
servers for a total of 203.04Ghz actual, so even without the multiplier
I'm not over provisioning my CPU Mhz.
Post by Yiping Zhang
Unless someone gives you better answer, I guess it's for fun - to
have more
Post by Yiping Zhang
detailed numbers in dashboard (may be it's related to other
hypervisor
Post by Yiping Zhang
types, just assuming... or not...)
Cheers
Post by Ugo Vasi
Hi Andrija,
not having noticed this new voice before I wondered if it is
limiting
Post by Yiping Zhang
Post by Ugo Vasi
the fact of reaching or exceeding the number of physical cores.
What is the purpose of this dashboard pane?
Post by Andrija Panic
Hi Ugo,
Why would you want to do this, just curious ?
I believe it's not possible, but anyway (at least with KVM,
probably same
Post by Yiping Zhang
Post by Ugo Vasi
Post by Andrija Panic
for other hypervisors) it doesn't even makes sense/use, since
when
Post by Yiping Zhang
Post by Ugo Vasi
Post by Andrija Panic
deploying a VM, ACS query host free/unused number of MHz (GHz),
so it's
Post by Yiping Zhang
Post by Ugo Vasi
not
Post by Andrija Panic
even relevant for ACS - number of cores in not relevant in ACS
calculations
Post by Andrija Panic
during VM deployment.
Cheers,
Andrija
Post by Ugo Vasi
Hi all,
in the dashboard of an ACS installation vesion 4.11.1.0
(Ubuntu 16.04
Post by Yiping Zhang
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
with KVM hypervisor), the new entry "# of CPU Cores" appears.
Is it possible to over-provision like for MHz or storage?
Thanks
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i
relativi
Post by Yiping Zhang
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento
trasmesso da parte
Post by Yiping Zhang
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
di qualsiasi soggetto diverso dal destinatario Ú proibita sia
ai sensi
Post by Yiping Zhang
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n.
196/2003
Post by Yiping Zhang
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
"Codice in materia di protezione dei dati personali". Se avete
ricevuto
Post by Yiping Zhang
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
questo messaggio per errore, vi preghiamo di distruggerlo e di
informare
Post by Yiping Zhang
Post by Ugo Vasi
Post by Andrija Panic
Post by Ugo Vasi
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i
relativi
Post by Yiping Zhang
Post by Ugo Vasi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso
da parte
Post by Yiping Zhang
Post by Ugo Vasi
di qualsiasi soggetto diverso dal destinatario Ú proibita sia ai
sensi
Post by Yiping Zhang
Post by Ugo Vasi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n.
196/2003
Post by Yiping Zhang
Post by Ugo Vasi
"Codice in materia di protezione dei dati personali". Se avete
ricevuto
Post by Yiping Zhang
Post by Ugo Vasi
questo messaggio per errore, vi preghiamo di distruggerlo e di
informare
Post by Yiping Zhang
Post by Ugo Vasi
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
--
Andrija Panić
--
Andrija Panić
Yiping Zhang
2018-11-20 00:29:35 UTC
Permalink
Eric:

If you change any of memory or cpu over provisioning factors, you need to restart (from GUI or API) all running VM instances, otherwise the reported allocation/available numbers would still reflect data before your changes. If you also add/delete instances at the same time, you could lose track of actual allocated and available resources very quickly until you restart all your instances.

Yiping
Post by Yiping Zhang
What's your value for global setting cpu.overprovisioning.factor?
I changed it to 4.0 and restarted the management server first thing. It
started out at 1.5. At 1.5, my zone shows 41% usage with the typical
workload -- 169.60Ghz / 409.78Ghz. I have 2x24x3.03ghz and 1x24x2.40ghz
servers for a total of 203.04Ghz actual, so even without the multiplier
I'm not over provisioning my CPU Mhz.
Post by Yiping Zhang
Unless someone gives you better answer, I guess it's for fun - to have more
detailed numbers in dashboard (may be it's related to other hypervisor
types, just assuming... or not...)
Cheers
Post by Ugo Vasi
Hi Andrija,
not having noticed this new voice before I wondered if it is limiting
the fact of reaching or exceeding the number of physical cores.
What is the purpose of this dashboard pane?
Post by Andrija Panic
Hi Ugo,
Why would you want to do this, just curious ?
I believe it's not possible, but anyway (at least with KVM, probably same
for other hypervisors) it doesn't even makes sense/use, since when
deploying a VM, ACS query host free/unused number of MHz (GHz), so it's
not
Post by Andrija Panic
even relevant for ACS - number of cores in not relevant in ACS
calculations
Post by Andrija Panic
during VM deployment.
Cheers,
Andrija
Post by Ugo Vasi
Hi all,
in the dashboard of an ACS installation vesion 4.11.1.0 (Ubuntu 16.04
with KVM hypervisor), the new entry "# of CPU Cores" appears.
Is it possible to over-provision like for MHz or storage?
Thanks
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da parte
di qualsiasi soggetto diverso dal destinatario è proibita sia ai sensi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete ricevuto
questo messaggio per errore, vi preghiamo di distruggerlo e di informare
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da parte
di qualsiasi soggetto diverso dal destinatario è proibita sia ai sensi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete ricevuto
questo messaggio per errore, vi preghiamo di distruggerlo e di informare
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
--
Andrij
Dag Sonstebo
2018-11-20 09:29:51 UTC
Permalink
This post might be inappropriate. Click to display it.
Eric Lee Green
2018-11-19 17:15:24 UTC
Permalink
Post by Andrija Panic
Hi Ugo,
Why would you want to do this, just curious ?
I believe it's not possible, but anyway (at least with KVM, probably same
for other hypervisors) it doesn't even makes sense/use, since when
deploying a VM, ACS query host free/unused number of MHz (GHz), so it's not
even relevant for ACS - number of cores in not relevant in ACS calculations
during VM deployment.
I think you are misunderstanding the question. I have 72 cores in my
cluster. Each of my hosts has 24 cores. With 4.9.2, I can provision 10
virtual machines, each of which is programmed with 8 cores, meaning 80
cores total. They on average are using only 25% of the CPU available to
them (they need to be able to burst) and my compute servers on average
are only 40% CPU used so that is not a problem.

When I tried upgrading to 4.11.1,  the dashboard showed a new value "#
of CPU Cores" in red and showed that I had more cores provisioned for
virtual machines than available in the cluster (80 versus 72 available).
Cloudstack would not launch new virtual machines. I shut down two
virtual machines, and now I can launch one, but not the second because I
would need 80 cores total in my cluster. I cannot launch all 10 virtual
machines because I would need 80 cores total. I know this because I
tried it. I then used MySQL to tell Cloudstack that each of my hosts has
48 cores (144 total), and suddenly I can launch all of my virtual machines.

Is this a bug in 4.11.1? Or is this expected behavior? If expected
behavior, is there a way to over-provision "total # of cores used" other
than to go into MySQL and tell it that my hosts have more cores than
they in fact have? (Note that my service offerings are limited to 8
cores max, so there's no way to launch a single VM with more cores than
exists on a physical host, since all my hosts have 24 cores).
Post by Andrija Panic
Post by Ugo Vasi
Hi all,
in the dashboard of an ACS installation vesion 4.11.1.0 (Ubuntu 16.04
with KVM hypervisor), the new entry "# of CPU Cores" appears.
Is it possible to over-provision like for MHz or storage?
Thanks
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da parte
di qualsiasi soggetto diverso dal destinatario è proibita sia ai sensi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete ricevuto
questo messaggio per errore, vi preghiamo di distruggerlo e di informare
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
Dag Sonstebo
2018-11-19 17:46:45 UTC
Permalink
Andrija - not sure about your 3.4GHz cores - must a be a simplified lookup somewhere making assumptions.

Eric - have just tried your scenario in my 4.11.2RC5 lab (admittedly with VMware, not KVM) - and I can see my core allocation keeps going up, e.g. at the moment it sits at 166% - 10 out of 6 cores used. However it doesn't stop me starting new VMs (only using 30-40% CPU and memory).
Are you sure it's the core count preventing you from starting VMs? What do the logs say? (Also keep in mind your system VMs are now using more resources that before).

Regards,
Dag Sonstebo
Cloud Architect
ShapeBlue
Post by Andrija Panic
Hi Ugo,
Why would you want to do this, just curious ?
I believe it's not possible, but anyway (at least with KVM, probably same
for other hypervisors) it doesn't even makes sense/use, since when
deploying a VM, ACS query host free/unused number of MHz (GHz), so it's not
even relevant for ACS - number of cores in not relevant in ACS calculations
during VM deployment.
I think you are misunderstanding the question. I have 72 cores in my
cluster. Each of my hosts has 24 cores. With 4.9.2, I can provision 10
virtual machines, each of which is programmed with 8 cores, meaning 80
cores total. They on average are using only 25% of the CPU available to
them (they need to be able to burst) and my compute servers on average
are only 40% CPU used so that is not a problem.

When I tried upgrading to 4.11.1, the dashboard showed a new value "#
of CPU Cores" in red and showed that I had more cores provisioned for
virtual machines than available in the cluster (80 versus 72 available).
Cloudstack would not launch new virtual machines. I shut down two
virtual machines, and now I can launch one, but not the second because I
would need 80 cores total in my cluster. I cannot launch all 10 virtual
machines because I would need 80 cores total. I know this because I
tried it. I then used MySQL to tell Cloudstack that each of my hosts has
48 cores (144 total), and suddenly I can launch all of my virtual machines.

Is this a bug in 4.11.1? Or is this expected behavior? If expected
behavior, is there a way to over-provision "total # of cores used" other
than to go into MySQL and tell it that my hosts have more cores than
they in fact have? (Note that my service offerings are limited to 8
cores max, so there's no way to launch a single VM with more cores than
exists on a physical host, since all my hosts have 24 cores).
Post by Andrija Panic
Post by Ugo Vasi
Hi all,
in the dashboard of an ACS installation vesion 4.11.1.0 (Ubuntu 16.04
with KVM hypervisor), the new entry "# of CPU Cores" appears.
Is it possible to over-provision like for MHz or storage?
Thanks
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da parte
di qualsiasi soggetto diverso dal destinatario è proibita sia ai sensi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete ricevuto
questo messaggio per errore, vi preghiamo di distruggerlo e di informare
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
***@shapeblue.com 
www.shapeblue.com
Amadeus House, Floral Street, London WC2E 9DPUK
@shap
Andrija Panic
2018-11-19 18:12:49 UTC
Permalink
Let me test this also, on clean 4.11.2 RC5... (KVM)
Post by Dag Sonstebo
Andrija - not sure about your 3.4GHz cores - must a be a simplified lookup
somewhere making assumptions.
Eric - have just tried your scenario in my 4.11.2RC5 lab (admittedly with
VMware, not KVM) - and I can see my core allocation keeps going up, e.g. at
the moment it sits at 166% - 10 out of 6 cores used. However it doesn't
stop me starting new VMs (only using 30-40% CPU and memory).
Are you sure it's the core count preventing you from starting VMs? What do
the logs say? (Also keep in mind your system VMs are now using more
resources that before).
Regards,
Dag Sonstebo
Cloud Architect
ShapeBlue
Post by Andrija Panic
Hi Ugo,
Why would you want to do this, just curious ?
I believe it's not possible, but anyway (at least with KVM, probably
same
Post by Andrija Panic
for other hypervisors) it doesn't even makes sense/use, since when
deploying a VM, ACS query host free/unused number of MHz (GHz), so
it's not
Post by Andrija Panic
even relevant for ACS - number of cores in not relevant in ACS
calculations
Post by Andrija Panic
during VM deployment.
I think you are misunderstanding the question. I have 72 cores in my
cluster. Each of my hosts has 24 cores. With 4.9.2, I can provision 10
virtual machines, each of which is programmed with 8 cores, meaning 80
cores total. They on average are using only 25% of the CPU available to
them (they need to be able to burst) and my compute servers on average
are only 40% CPU used so that is not a problem.
When I tried upgrading to 4.11.1, the dashboard showed a new value "#
of CPU Cores" in red and showed that I had more cores provisioned for
virtual machines than available in the cluster (80 versus 72 available).
Cloudstack would not launch new virtual machines. I shut down two
virtual machines, and now I can launch one, but not the second because I
would need 80 cores total in my cluster. I cannot launch all 10 virtual
machines because I would need 80 cores total. I know this because I
tried it. I then used MySQL to tell Cloudstack that each of my hosts has
48 cores (144 total), and suddenly I can launch all of my virtual machines.
Is this a bug in 4.11.1? Or is this expected behavior? If expected
behavior, is there a way to over-provision "total # of cores used" other
than to go into MySQL and tell it that my hosts have more cores than
they in fact have? (Note that my service offerings are limited to 8
cores max, so there's no way to launch a single VM with more cores than
exists on a physical host, since all my hosts have 24 cores).
Post by Andrija Panic
Post by Ugo Vasi
Hi all,
in the dashboard of an ACS installation vesion 4.11.1.0 (Ubuntu
16.04
Post by Andrija Panic
Post by Ugo Vasi
with KVM hypervisor), the new entry "# of CPU Cores" appears.
Is it possible to over-provision like for MHz or storage?
Thanks
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da
parte
Post by Andrija Panic
Post by Ugo Vasi
di qualsiasi soggetto diverso dal destinatario Ú proibita sia ai
sensi
Post by Andrija Panic
Post by Ugo Vasi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete
ricevuto
Post by Andrija Panic
Post by Ugo Vasi
questo messaggio per errore, vi preghiamo di distruggerlo e di
informare
Post by Andrija Panic
Post by Ugo Vasi
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
www.shapeblue.com
Amadeus House, Floral Street, London WC2E 9DPUK
@shapeblue
--
Andrija Panić
Andrija Panic
2018-11-19 18:25:49 UTC
Permalink
Loading Image...

All fine, but note that my CPU GHZ is NOT over 80% (I believe 0.8 aka 80%
is the cpu warning or disable threshold)

As Dag said, please check from mgmt logs why you can't actually spin new VM
in your case.

When you start to deploy VM, there will be some friendly printed (human
readable) lines similar to "checking list of hosts with enough resources"
and similar, it will evaluate each host for needed number of MHz and RAM
memory - and if none match, then it will fail to deploy - and it will say
that clearly, something like "no host with enough CPU capacity" or in
similar fashion...

(this is clean install though...)

Please check your "cluster.cpu.allocated.capacity.disablethreshold" both in
Global Settings, and on particular Cluster level.

Cheers
Post by Andrija Panic
Let me test this also, on clean 4.11.2 RC5... (KVM)
Post by Dag Sonstebo
Andrija - not sure about your 3.4GHz cores - must a be a simplified
lookup somewhere making assumptions.
Eric - have just tried your scenario in my 4.11.2RC5 lab (admittedly with
VMware, not KVM) - and I can see my core allocation keeps going up, e.g. at
the moment it sits at 166% - 10 out of 6 cores used. However it doesn't
stop me starting new VMs (only using 30-40% CPU and memory).
Are you sure it's the core count preventing you from starting VMs? What
do the logs say? (Also keep in mind your system VMs are now using more
resources that before).
Regards,
Dag Sonstebo
Cloud Architect
ShapeBlue
Post by Andrija Panic
Hi Ugo,
Why would you want to do this, just curious ?
I believe it's not possible, but anyway (at least with KVM,
probably same
Post by Andrija Panic
for other hypervisors) it doesn't even makes sense/use, since when
deploying a VM, ACS query host free/unused number of MHz (GHz), so
it's not
Post by Andrija Panic
even relevant for ACS - number of cores in not relevant in ACS
calculations
Post by Andrija Panic
during VM deployment.
I think you are misunderstanding the question. I have 72 cores in my
cluster. Each of my hosts has 24 cores. With 4.9.2, I can provision 10
virtual machines, each of which is programmed with 8 cores, meaning 80
cores total. They on average are using only 25% of the CPU available to
them (they need to be able to burst) and my compute servers on average
are only 40% CPU used so that is not a problem.
When I tried upgrading to 4.11.1, the dashboard showed a new value "#
of CPU Cores" in red and showed that I had more cores provisioned for
virtual machines than available in the cluster (80 versus 72 available).
Cloudstack would not launch new virtual machines. I shut down two
virtual machines, and now I can launch one, but not the second because I
would need 80 cores total in my cluster. I cannot launch all 10 virtual
machines because I would need 80 cores total. I know this because I
tried it. I then used MySQL to tell Cloudstack that each of my hosts has
48 cores (144 total), and suddenly I can launch all of my virtual machines.
Is this a bug in 4.11.1? Or is this expected behavior? If expected
behavior, is there a way to over-provision "total # of cores used" other
than to go into MySQL and tell it that my hosts have more cores than
they in fact have? (Note that my service offerings are limited to 8
cores max, so there's no way to launch a single VM with more cores than
exists on a physical host, since all my hosts have 24 cores).
Post by Andrija Panic
Post by Ugo Vasi
Hi all,
in the dashboard of an ACS installation vesion 4.11.1.0 (Ubuntu
16.04
Post by Andrija Panic
Post by Ugo Vasi
with KVM hypervisor), the new entry "# of CPU Cores" appears.
Is it possible to over-provision like for MHz or storage?
Thanks
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i
relativi
Post by Andrija Panic
Post by Ugo Vasi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da
parte
Post by Andrija Panic
Post by Ugo Vasi
di qualsiasi soggetto diverso dal destinatario Ú proibita sia ai
sensi
Post by Andrija Panic
Post by Ugo Vasi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n.
196/2003
Post by Andrija Panic
Post by Ugo Vasi
"Codice in materia di protezione dei dati personali". Se avete
ricevuto
Post by Andrija Panic
Post by Ugo Vasi
questo messaggio per errore, vi preghiamo di distruggerlo e di
informare
Post by Andrija Panic
Post by Ugo Vasi
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
www.shapeblue.com
Amadeus House, Floral Street, London WC2E 9DPUK
@shapeblue
--
Andrija Panić
--
Andrija Panić
Eric Green
2018-11-19 19:42:24 UTC
Permalink
This post might be inappropriate. Click to display it.
Andrija Panic
2018-11-19 20:16:25 UTC
Permalink
Eric, I sent you numbers from KVM, Ubuntu 14.04 (qemu/libvrti from Ubuntu
16.04)

Can you please reproduce the issue and upload logs - make sure to truncate
logs prior to clicking the "FINISH" button in the VM creation wizard..
Upload to pastebin, so we can check it

Cheers
Post by Eric Green
I am with KVM.
I am sure it’s the core count preventing me from starting VM’s because
when I hack the database to tell it I have 48 cores rather than 24 cores on
my hosts, I can then start the VM.
The only thing the logs say is that I can’t create a new VM due to lack of
resources. Then it quits saying that when I hack the database. Note that
under 4.9.2 (what I reverted back to), Memory is at 49%, CPU is at 41%,
Primary Storage is at 5%, and Secondary Storage is at 5%. All other
resources aren’t even 1% used (I set up # of vlans, shared network IP’s,
etc. fairly large because I expect to grow the cluster in the future). 4.9
doesn’t list CPU cores. Under 4.11.1 those measures were the same.
I am running KVM under Centos 7. It may be that the KVM allocator works
different from the VMware allocator?
From: Dag Sonstebo
Sent: Monday, November 19, 2018 9:47 AM
Subject: Re: number of cores
Andrija - not sure about your 3.4GHz cores - must a be a simplified lookup
somewhere making assumptions.
Eric - have just tried your scenario in my 4.11.2RC5 lab (admittedly with
VMware, not KVM) - and I can see my core allocation keeps going up, e.g. at
the moment it sits at 166% - 10 out of 6 cores used. However it doesn't
stop me starting new VMs (only using 30-40% CPU and memory).
Are you sure it's the core count preventing you from starting VMs? What do
the logs say? (Also keep in mind your system VMs are now using more
resources that before).
Regards,
Dag Sonstebo
Cloud Architect
ShapeBlue
Post by Andrija Panic
Hi Ugo,
Why would you want to do this, just curious ?
I believe it's not possible, but anyway (at least with KVM, probably
same
Post by Andrija Panic
for other hypervisors) it doesn't even makes sense/use, since when
deploying a VM, ACS query host free/unused number of MHz (GHz), so
it's not
Post by Andrija Panic
even relevant for ACS - number of cores in not relevant in ACS
calculations
Post by Andrija Panic
during VM deployment.
I think you are misunderstanding the question. I have 72 cores in my
cluster. Each of my hosts has 24 cores. With 4.9.2, I can provision 10
virtual machines, each of which is programmed with 8 cores, meaning 80
cores total. They on average are using only 25% of the CPU available to
them (they need to be able to burst) and my compute servers on average
are only 40% CPU used so that is not a problem.
When I tried upgrading to 4.11.1, the dashboard showed a new value "#
of CPU Cores" in red and showed that I had more cores provisioned for
virtual machines than available in the cluster (80 versus 72 available).
Cloudstack would not launch new virtual machines. I shut down two
virtual machines, and now I can launch one, but not the second because I
would need 80 cores total in my cluster. I cannot launch all 10 virtual
machines because I would need 80 cores total. I know this because I
tried it. I then used MySQL to tell Cloudstack that each of my hosts has
48 cores (144 total), and suddenly I can launch all of my virtual machines.
Is this a bug in 4.11.1? Or is this expected behavior? If expected
behavior, is there a way to over-provision "total # of cores used" other
than to go into MySQL and tell it that my hosts have more cores than
they in fact have? (Note that my service offerings are limited to 8
cores max, so there's no way to launch a single VM with more cores than
exists on a physical host, since all my hosts have 24 cores).
Post by Andrija Panic
Post by Ugo Vasi
Hi all,
in the dashboard of an ACS installation vesion 4.11.1.0 (Ubuntu
16.04
Post by Andrija Panic
Post by Ugo Vasi
with KVM hypervisor), the new entry "# of CPU Cores" appears.
Is it possible to over-provision like for MHz or storage?
Thanks
--
*Ugo Vasi* / System Administrator
*Procne S.r.l.*
+39 0432 486 523
via Cotonificio, 45
33010 Tavagnacco (UD)
www.procne.it <http://www.procne.it/>
Le informazioni contenute nella presente comunicazione ed i relativi
allegati possono essere riservate e sono, comunque, destinate
esclusivamente alle persone od alla Società sopraindicati. La
diffusione, distribuzione e/o copiatura del documento trasmesso da
parte
Post by Andrija Panic
Post by Ugo Vasi
di qualsiasi soggetto diverso dal destinatario Ú proibita sia ai
sensi
Post by Andrija Panic
Post by Ugo Vasi
dell'art. 616 c.p., che ai sensi del Decreto Legislativo n. 196/2003
"Codice in materia di protezione dei dati personali". Se avete
ricevuto
Post by Andrija Panic
Post by Ugo Vasi
questo messaggio per errore, vi preghiamo di distruggerlo e di
informare
Post by Andrija Panic
Post by Ugo Vasi
immediatamente Procne S.r.l. scrivendo all' indirizzo e-mail
www.shapeblue.com
Amadeus House, Floral Street, London WC2E 9DPUK
@shapeblue
--
Andrija Panić
Loading...