Login | Register For Free | Help
Search for: (Advanced)

Mailing List Archive: Linux Virtual Server: Users

[lvs-users] Question on SH scheduler

 

 

Linux Virtual Server users RSS feed   Index | Next | Previous | View Threaded


krishna.sirigiri at gmail

Feb 27, 2012, 5:08 AM

Post #1 of 13 (909 views)
Permalink
[lvs-users] Question on SH scheduler

Dear All,
If the director is configured with Source Hashing scheduling algorithm,
the connections are distributed among the back end real servers through
looking up a statically assigned hash table by their source IP addresses.
My question here is the key for the load balancing is the source IP alone
or source IP plus port?

Any help on this is highly appreciated.

Thanks
Krishna Prasad.
_______________________________________________
Please read the documentation before posting - it's available at:
http://www.linuxvirtualserver.org/

LinuxVirtualServer.org mailing list - lvs-users [at] LinuxVirtualServer
Send requests to lvs-users-request [at] LinuxVirtualServer
or go to http://lists.graemef.net/mailman/listinfo/lvs-users


khosrow.ebrahimpour at ec

Feb 27, 2012, 7:32 AM

Post #2 of 13 (866 views)
Permalink
Re: [lvs-users] Question on SH scheduler [In reply to]

Hi Krishna,

On Monday, February 27, 2012 08:08:13 am krishna prasad wrote:
> Dear All,
> If the director is configured with Source Hashing scheduling algorithm,
> the connections are distributed among the back end real servers through
> looking up a statically assigned hash table by their source IP addresses.
> My question here is the key for the load balancing is the source IP alone
> or source IP plus port?

I haven't used SH scheduling myself, but according to this article
http://kb.linuxvirtualserver.org/wiki/Source_Hashing_Scheduling only the
source IP is used.

Although I am curious why the hash wouldn't use source IP + Port. Otherwise,
clients coming from behind a NAT or proxy will all end up on the same
realserver.

--
Khosrow

_______________________________________________
Please read the documentation before posting - it's available at:
http://www.linuxvirtualserver.org/

LinuxVirtualServer.org mailing list - lvs-users [at] LinuxVirtualServer
Send requests to lvs-users-request [at] LinuxVirtualServer
or go to http://lists.graemef.net/mailman/listinfo/lvs-users


horms at verge

Feb 27, 2012, 2:48 PM

Post #3 of 13 (878 views)
Permalink
Re: [lvs-users] Question on SH scheduler [In reply to]

On Mon, Feb 27, 2012 at 10:32:16AM -0500, Khosrow Ebrahimpour wrote:
> Hi Krishna,
>
> On Monday, February 27, 2012 08:08:13 am krishna prasad wrote:
> > Dear All,
> > If the director is configured with Source Hashing scheduling algorithm,
> > the connections are distributed among the back end real servers through
> > looking up a statically assigned hash table by their source IP addresses.
> > My question here is the key for the load balancing is the source IP alone
> > or source IP plus port?
>
> I haven't used SH scheduling myself, but according to this article
> http://kb.linuxvirtualserver.org/wiki/Source_Hashing_Scheduling only the
> source IP is used.

Yes, the SH scheduler only makes use of the source IP address and
conversely the DH scheduler only makes use of the destination IP address.

> Although I am curious why the hash wouldn't use source IP + Port. Otherwise,
> clients coming from behind a NAT or proxy will all end up on the same
> realserver.

I believe that the motivation for the DH scheduler was for use with
load-balanced caching proxy servers. The SH scheduler is intended
to be used in place of persistence in some situations where it is
desirable to scheduler the same client to the same real-server.

Clearly the presence of NAT can potentially result in a poor result
when using SH. Likewise with persistence, which can provide some
of the same behaviour.

With regards to SH hashing on both the Source IP and port, I'm unclear
of when this would be useful in place of for example WLC



_______________________________________________
Please read the documentation before posting - it's available at:
http://www.linuxvirtualserver.org/

LinuxVirtualServer.org mailing list - lvs-users [at] LinuxVirtualServer
Send requests to lvs-users-request [at] LinuxVirtualServer
or go to http://lists.graemef.net/mailman/listinfo/lvs-users


krishna.sirigiri at gmail

Feb 27, 2012, 7:26 PM

Post #4 of 13 (860 views)
Permalink
Re: [lvs-users] Question on SH scheduler [In reply to]

On Tue, Feb 28, 2012 at 4:18 AM, Simon Horman <horms [at] verge> wrote:

> On Mon, Feb 27, 2012 at 10:32:16AM -0500, Khosrow Ebrahimpour wrote:
> > Hi Krishna,
> >
> > On Monday, February 27, 2012 08:08:13 am krishna prasad wrote:
> > > Dear All,
> > > If the director is configured with Source Hashing scheduling
> algorithm,
> > > the connections are distributed among the back end real servers through
> > > looking up a statically assigned hash table by their source IP
> addresses.
> > > My question here is the key for the load balancing is the source IP
> alone
> > > or source IP plus port?
> >
> > I haven't used SH scheduling myself, but according to this article
> > http://kb.linuxvirtualserver.org/wiki/Source_Hashing_Scheduling only the
> > source IP is used.
>
> Yes, the SH scheduler only makes use of the source IP address and
> conversely the DH scheduler only makes use of the destination IP address.
>
> > Although I am curious why the hash wouldn't use source IP + Port.
> Otherwise,
> > clients coming from behind a NAT or proxy will all end up on the same
> > realserver.
>
> I believe that the motivation for the DH scheduler was for use with
> load-balanced caching proxy servers. The SH scheduler is intended
> to be used in place of persistence in some situations where it is
> desirable to scheduler the same client to the same real-server.
>
> Clearly the presence of NAT can potentially result in a poor result
> when using SH. Likewise with persistence, which can provide some
> of the same behaviour.
>
> With regards to SH hashing on both the Source IP and port, I'm unclear
> of when this would be useful in place of for example WLC
>
> I agree that the original motivations of SH (and DH) may not needed to
> have IP+port hash;
>
But I strongly think that it it good to have IP+port hashing, for cases
where multiple clients run on single host, in this case
the connections have same IP but different port. In this case also the same
is desirable,i.e same client to the same real-server.
This may not make a real use case for web world, but a strong case for
non-web deployments like in telecom.I know LVS is increasingly used in
other than web services.



>
> _______________________________________________
> Please read the documentation before posting - it's available at:
> http://www.linuxvirtualserver.org/
>
> LinuxVirtualServer.org mailing list - lvs-users [at] LinuxVirtualServer
> Send requests to lvs-users-request [at] LinuxVirtualServer
> or go to http://lists.graemef.net/mailman/listinfo/lvs-users
>
_______________________________________________
Please read the documentation before posting - it's available at:
http://www.linuxvirtualserver.org/

LinuxVirtualServer.org mailing list - lvs-users [at] LinuxVirtualServer
Send requests to lvs-users-request [at] LinuxVirtualServer
or go to http://lists.graemef.net/mailman/listinfo/lvs-users


krishna.sirigiri at gmail

Feb 27, 2012, 8:05 PM

Post #5 of 13 (868 views)
Permalink
Re: [lvs-users] Question on SH scheduler [In reply to]

By the way, is there any other scheduler (other than SH) by which we can
deterministically forward the client connections to the back end real
servers?
I mean connection from client1 should go to real server 1..etc.

On Tue, Feb 28, 2012 at 8:56 AM, krishna prasad
<krishna.sirigiri [at] gmail>wrote:

>
>
> On Tue, Feb 28, 2012 at 4:18 AM, Simon Horman <horms [at] verge> wrote:
>
>> On Mon, Feb 27, 2012 at 10:32:16AM -0500, Khosrow Ebrahimpour wrote:
>> > Hi Krishna,
>> >
>> > On Monday, February 27, 2012 08:08:13 am krishna prasad wrote:
>> > > Dear All,
>> > > If the director is configured with Source Hashing scheduling
>> algorithm,
>> > > the connections are distributed among the back end real servers
>> through
>> > > looking up a statically assigned hash table by their source IP
>> addresses.
>> > > My question here is the key for the load balancing is the source IP
>> alone
>> > > or source IP plus port?
>> >
>> > I haven't used SH scheduling myself, but according to this article
>> > http://kb.linuxvirtualserver.org/wiki/Source_Hashing_Scheduling only
>> the
>> > source IP is used.
>>
>> Yes, the SH scheduler only makes use of the source IP address and
>> conversely the DH scheduler only makes use of the destination IP address.
>>
>> > Although I am curious why the hash wouldn't use source IP + Port.
>> Otherwise,
>> > clients coming from behind a NAT or proxy will all end up on the same
>> > realserver.
>>
>> I believe that the motivation for the DH scheduler was for use with
>> load-balanced caching proxy servers. The SH scheduler is intended
>> to be used in place of persistence in some situations where it is
>> desirable to scheduler the same client to the same real-server.
>>
>> Clearly the presence of NAT can potentially result in a poor result
>> when using SH. Likewise with persistence, which can provide some
>> of the same behaviour.
>>
>> With regards to SH hashing on both the Source IP and port, I'm unclear
>> of when this would be useful in place of for example WLC
>>
>> I agree that the original motivations of SH (and DH) may not needed to
>> have IP+port hash;
>>
> But I strongly think that it it good to have IP+port hashing, for cases
> where multiple clients run on single host, in this case
> the connections have same IP but different port. In this case also the
> same is desirable,i.e same client to the same real-server.
> This may not make a real use case for web world, but a strong case for
> non-web deployments like in telecom.I know LVS is increasingly used in
> other than web services.
>
>
>
>>
>> _______________________________________________
>> Please read the documentation before posting - it's available at:
>> http://www.linuxvirtualserver.org/
>>
>> LinuxVirtualServer.org mailing list - lvs-users [at] LinuxVirtualServer
>> Send requests to lvs-users-request [at] LinuxVirtualServer
>> or go to http://lists.graemef.net/mailman/listinfo/lvs-users
>>
>
>
_______________________________________________
Please read the documentation before posting - it's available at:
http://www.linuxvirtualserver.org/

LinuxVirtualServer.org mailing list - lvs-users [at] LinuxVirtualServer
Send requests to lvs-users-request [at] LinuxVirtualServer
or go to http://lists.graemef.net/mailman/listinfo/lvs-users


holler at ahsoftware

Feb 27, 2012, 8:34 PM

Post #6 of 13 (864 views)
Permalink
Re: [lvs-users] Question on SH scheduler [In reply to]

Hello,

Am 28.02.2012 04:26, schrieb krishna prasad:

> But I strongly think that it it good to have IP+port hashing, for cases
> where multiple clients run on single host, in this case
> the connections have same IP but different port. In this case also the same
> is desirable,i.e same client to the same real-server.
> This may not make a real use case for web world, but a strong case for
> non-web deployments like in telecom.I know LVS is increasingly used in
> other than web services.

What should be the use case for this? Source ports are almost always
choosen randomly, so you woould get the same results as balancing randomly.

Regards,

Alexander

_______________________________________________
Please read the documentation before posting - it's available at:
http://www.linuxvirtualserver.org/

LinuxVirtualServer.org mailing list - lvs-users [at] LinuxVirtualServer
Send requests to lvs-users-request [at] LinuxVirtualServer
or go to http://lists.graemef.net/mailman/listinfo/lvs-users


horms at verge

Feb 27, 2012, 8:54 PM

Post #7 of 13 (875 views)
Permalink
Re: [lvs-users] Question on SH scheduler [In reply to]

On Tue, Feb 28, 2012 at 09:35:23AM +0530, krishna prasad wrote:
> By the way, is there any other scheduler (other than SH) by which we can
> deterministically forward the client connections to the back end real
> servers?
> I mean connection from client1 should go to real server 1..etc.

I took a quick look this morning, and I don't think there is such a
scheduler.

FWIW, it should be easy enough to implement your idea by
copying the SH scheduler and adding the port to the hash
in the new code.

_______________________________________________
Please read the documentation before posting - it's available at:
http://www.linuxvirtualserver.org/

LinuxVirtualServer.org mailing list - lvs-users [at] LinuxVirtualServer
Send requests to lvs-users-request [at] LinuxVirtualServer
or go to http://lists.graemef.net/mailman/listinfo/lvs-users


krishna.sirigiri at gmail

Feb 27, 2012, 10:06 PM

Post #8 of 13 (859 views)
Permalink
Re: [lvs-users] Question on SH scheduler [In reply to]

On Tue, Feb 28, 2012 at 11:18 AM, Alexander Holler <holler [at] ahsoftware>wrote:

> Am 28.02.2012 06:35, schrieb krishna prasad:
>
>> On Tue, Feb 28, 2012 at 10:44 AM, Alexander Holler<holler [at] ahsoftware>*
>> *wrote:
>>
>> Am 28.02.2012 06:11, schrieb krishna prasad:
>>>
>>> On Tue, Feb 28, 2012 at 10:04 AM, Alexander Holler<holler [at] ahsoftware
>>> >*
>>>
>>>> *wrote:
>>>>
>>>>
>>>> Hello,
>>>>
>>>>>
>>>>> Am 28.02.2012 04:26, schrieb krishna prasad:
>>>>>
>>>>>
>>>>> But I strongly think that it it good to have IP+port hashing, for
>>>>> cases
>>>>>
>>>>> where multiple clients run on single host, in this case
>>>>>> the connections have same IP but different port. In this case also the
>>>>>> same
>>>>>> is desirable,i.e same client to the same real-server.
>>>>>> This may not make a real use case for web world, but a strong case for
>>>>>> non-web deployments like in telecom.I know LVS is increasingly used in
>>>>>> other than web services.
>>>>>>
>>>>>>
>>>>>> What should be the use case for this? Source ports are almost always
>>>>> choosen randomly, so you woould get the same results as balancing
>>>>> randomly.
>>>>>
>>>>>
>>>>> Not necessarily,I came across few implementations where client port is
>>>> fixed (they bind() port while creating socket), but I agree that most of
>>>>
>>>>
>>> Sure.
>>>
>>>
>>> the times source port is random.
>>>
>>>> The good approach would be LVS to provide options for IP+port or just
>>>> IP
>>>> hashing.
>>>>
>>>>
>>> But I still miss the use case. If the client always does come with the
>>> same port, it doesn't make a difference if the port is used too for
>>> hashing
>>> or just the IP.
>>>
>>>
>> Oh..Sorry, I should have been clear.
>> Here is the use case: I have a client (from out side it looks like a piza
>> box, but internally it has many CPU..something like blade server/ATCA)
>> which initiates TCP/SCTP connections with same IP address but with
>> different Port.
>> So if I use SH, all these connections (potentially this client can
>> initiate
>> as many as 40 connections) will land on a same real server which may not
>> be
>> what we wanted. we wanted the connections to be balanced (based on
>> IP+port)
>> across all the real servers. Does it make sense?
>>
>
> Hmm, if it comes to a few thousand different ports, it would, but not for
> 40 (imho). ;)
>
> Anyway, this discussion should be held at the ML, not private. Maybe
> someone else could have add some ideas.
>

Oh..did not notice this, this time I am including the ML group.

>
> Regards,
>
> Alexander
>
_______________________________________________
Please read the documentation before posting - it's available at:
http://www.linuxvirtualserver.org/

LinuxVirtualServer.org mailing list - lvs-users [at] LinuxVirtualServer
Send requests to lvs-users-request [at] LinuxVirtualServer
or go to http://lists.graemef.net/mailman/listinfo/lvs-users


horms at verge

Feb 27, 2012, 11:12 PM

Post #9 of 13 (855 views)
Permalink
Re: [lvs-users] Question on SH scheduler [In reply to]

On Tue, Feb 28, 2012 at 11:36:11AM +0530, krishna prasad wrote:
> On Tue, Feb 28, 2012 at 11:18 AM, Alexander Holler <holler [at] ahsoftware>wrote:
>
> > Am 28.02.2012 06:35, schrieb krishna prasad:
> >
> >> On Tue, Feb 28, 2012 at 10:44 AM, Alexander Holler<holler [at] ahsoftware>*
> >> *wrote:
> >>
> >> Am 28.02.2012 06:11, schrieb krishna prasad:
> >>>
> >>> On Tue, Feb 28, 2012 at 10:04 AM, Alexander Holler<holler [at] ahsoftware
> >>> >*
> >>>
> >>>> *wrote:
> >>>>
> >>>>
> >>>> Hello,
> >>>>
> >>>>>
> >>>>> Am 28.02.2012 04:26, schrieb krishna prasad:
> >>>>>
> >>>>>
> >>>>> But I strongly think that it it good to have IP+port hashing, for
> >>>>> cases
> >>>>>
> >>>>> where multiple clients run on single host, in this case
> >>>>>> the connections have same IP but different port. In this case also the
> >>>>>> same
> >>>>>> is desirable,i.e same client to the same real-server.
> >>>>>> This may not make a real use case for web world, but a strong case for
> >>>>>> non-web deployments like in telecom.I know LVS is increasingly used in
> >>>>>> other than web services.
> >>>>>>
> >>>>>>
> >>>>>> What should be the use case for this? Source ports are almost always
> >>>>> choosen randomly, so you woould get the same results as balancing
> >>>>> randomly.
> >>>>>
> >>>>>
> >>>>> Not necessarily,I came across few implementations where client port is
> >>>> fixed (they bind() port while creating socket), but I agree that most of
> >>>>
> >>>>
> >>> Sure.
> >>>
> >>>
> >>> the times source port is random.
> >>>
> >>>> The good approach would be LVS to provide options for IP+port or just
> >>>> IP
> >>>> hashing.
> >>>>
> >>>>
> >>> But I still miss the use case. If the client always does come with the
> >>> same port, it doesn't make a difference if the port is used too for
> >>> hashing
> >>> or just the IP.
> >>>
> >>>
> >> Oh..Sorry, I should have been clear.
> >> Here is the use case: I have a client (from out side it looks like a piza
> >> box, but internally it has many CPU..something like blade server/ATCA)
> >> which initiates TCP/SCTP connections with same IP address but with
> >> different Port.
> >> So if I use SH, all these connections (potentially this client can
> >> initiate
> >> as many as 40 connections) will land on a same real server which may not
> >> be
> >> what we wanted. we wanted the connections to be balanced (based on
> >> IP+port)
> >> across all the real servers. Does it make sense?
> >>
> >
> > Hmm, if it comes to a few thousand different ports, it would, but not for
> > 40 (imho). ;)
> >
> > Anyway, this discussion should be held at the ML, not private. Maybe
> > someone else could have add some ideas.
> >
>
> Oh..did not notice this, this time I am including the ML group.

Hi,

I think that the scenario described above does make sense and that
none of the existing schedulers cater to it. Although persistence could be
used to achieve much if not all of the desired result. I would be happy
to consider a new scheduler that implements source address+port hashing.


_______________________________________________
Please read the documentation before posting - it's available at:
http://www.linuxvirtualserver.org/

LinuxVirtualServer.org mailing list - lvs-users [at] LinuxVirtualServer
Send requests to lvs-users-request [at] LinuxVirtualServer
or go to http://lists.graemef.net/mailman/listinfo/lvs-users


krishna.sirigiri at gmail

Feb 28, 2012, 12:36 AM

Post #10 of 13 (861 views)
Permalink
Re: [lvs-users] Question on SH scheduler [In reply to]

Hi Simon,

On Tue, Feb 28, 2012 at 12:42 PM, Simon Horman <horms [at] verge> wrote:

> On Tue, Feb 28, 2012 at 11:36:11AM +0530, krishna prasad wrote:
> > On Tue, Feb 28, 2012 at 11:18 AM, Alexander Holler <holler [at] ahsoftware
> >wrote:
> >
> > > Am 28.02.2012 06:35, schrieb krishna prasad:
> > >
> > >> On Tue, Feb 28, 2012 at 10:44 AM, Alexander Holler<
> holler [at] ahsoftware>*
> > >> *wrote:
> > >>
> > >> Am 28.02.2012 06:11, schrieb krishna prasad:
> > >>>
> > >>> On Tue, Feb 28, 2012 at 10:04 AM, Alexander Holler<
> holler [at] ahsoftware
> > >>> >*
> > >>>
> > >>>> *wrote:
> > >>>>
> > >>>>
> > >>>> Hello,
> > >>>>
> > >>>>>
> > >>>>> Am 28.02.2012 04:26, schrieb krishna prasad:
> > >>>>>
> > >>>>>
> > >>>>> But I strongly think that it it good to have IP+port hashing,
> for
> > >>>>> cases
> > >>>>>
> > >>>>> where multiple clients run on single host, in this case
> > >>>>>> the connections have same IP but different port. In this case
> also the
> > >>>>>> same
> > >>>>>> is desirable,i.e same client to the same real-server.
> > >>>>>> This may not make a real use case for web world, but a strong
> case for
> > >>>>>> non-web deployments like in telecom.I know LVS is increasingly
> used in
> > >>>>>> other than web services.
> > >>>>>>
> > >>>>>>
> > >>>>>> What should be the use case for this? Source ports are almost
> always
> > >>>>> choosen randomly, so you woould get the same results as balancing
> > >>>>> randomly.
> > >>>>>
> > >>>>>
> > >>>>> Not necessarily,I came across few implementations where client
> port is
> > >>>> fixed (they bind() port while creating socket), but I agree that
> most of
> > >>>>
> > >>>>
> > >>> Sure.
> > >>>
> > >>>
> > >>> the times source port is random.
> > >>>
> > >>>> The good approach would be LVS to provide options for IP+port or
> just
> > >>>> IP
> > >>>> hashing.
> > >>>>
> > >>>>
> > >>> But I still miss the use case. If the client always does come with
> the
> > >>> same port, it doesn't make a difference if the port is used too for
> > >>> hashing
> > >>> or just the IP.
> > >>>
> > >>>
> > >> Oh..Sorry, I should have been clear.
> > >> Here is the use case: I have a client (from out side it looks like a
> piza
> > >> box, but internally it has many CPU..something like blade server/ATCA)
> > >> which initiates TCP/SCTP connections with same IP address but with
> > >> different Port.
> > >> So if I use SH, all these connections (potentially this client can
> > >> initiate
> > >> as many as 40 connections) will land on a same real server which may
> not
> > >> be
> > >> what we wanted. we wanted the connections to be balanced (based on
> > >> IP+port)
> > >> across all the real servers. Does it make sense?
> > >>
> > >
> > > Hmm, if it comes to a few thousand different ports, it would, but not
> for
> > > 40 (imho). ;)
> > >
> > > Anyway, this discussion should be held at the ML, not private. Maybe
> > > someone else could have add some ideas.
> > >
> >
> > Oh..did not notice this, this time I am including the ML group.
>
> Hi,
>
> I think that the scenario described above does make sense and that
> none of the existing schedulers cater to it. Although persistence could be
> used to achieve much if not all of the desired result. I would be happy
> to consider a new scheduler that implements source address+port hashing.
>
> the persistence mentioned above, how does it work?Can you point to me a
> link or tutorial on this?
>
Also, do you have any idea if some one already working on this kind of new
schedule?


_______________________________________________
> Please read the documentation before posting - it's available at:
> http://www.linuxvirtualserver.org/
>
> LinuxVirtualServer.org mailing list - lvs-users [at] LinuxVirtualServer
> Send requests to lvs-users-request [at] LinuxVirtualServer
> or go to http://lists.graemef.net/mailman/listinfo/lvs-users
>
_______________________________________________
Please read the documentation before posting - it's available at:
http://www.linuxvirtualserver.org/

LinuxVirtualServer.org mailing list - lvs-users [at] LinuxVirtualServer
Send requests to lvs-users-request [at] LinuxVirtualServer
or go to http://lists.graemef.net/mailman/listinfo/lvs-users


horms at verge

Feb 28, 2012, 1:35 AM

Post #11 of 13 (859 views)
Permalink
Re: [lvs-users] Question on SH scheduler [In reply to]

On Tue, Feb 28, 2012 at 02:06:12PM +0530, krishna prasad wrote:
> Hi Simon,
>
> On Tue, Feb 28, 2012 at 12:42 PM, Simon Horman <horms [at] verge> wrote:
>
> > On Tue, Feb 28, 2012 at 11:36:11AM +0530, krishna prasad wrote:
> > > On Tue, Feb 28, 2012 at 11:18 AM, Alexander Holler <holler [at] ahsoftware
> > >wrote:
> > >
> > > > Am 28.02.2012 06:35, schrieb krishna prasad:
> > > >
> > > >> On Tue, Feb 28, 2012 at 10:44 AM, Alexander Holler<
> > holler [at] ahsoftware>*
> > > >> *wrote:
> > > >>
> > > >> Am 28.02.2012 06:11, schrieb krishna prasad:
> > > >>>
> > > >>> On Tue, Feb 28, 2012 at 10:04 AM, Alexander Holler<
> > holler [at] ahsoftware
> > > >>> >*
> > > >>>
> > > >>>> *wrote:
> > > >>>>
> > > >>>>
> > > >>>> Hello,
> > > >>>>
> > > >>>>>
> > > >>>>> Am 28.02.2012 04:26, schrieb krishna prasad:
> > > >>>>>
> > > >>>>>
> > > >>>>> But I strongly think that it it good to have IP+port hashing,
> > for
> > > >>>>> cases
> > > >>>>>
> > > >>>>> where multiple clients run on single host, in this case
> > > >>>>>> the connections have same IP but different port. In this case
> > also the
> > > >>>>>> same
> > > >>>>>> is desirable,i.e same client to the same real-server.
> > > >>>>>> This may not make a real use case for web world, but a strong
> > case for
> > > >>>>>> non-web deployments like in telecom.I know LVS is increasingly
> > used in
> > > >>>>>> other than web services.
> > > >>>>>>
> > > >>>>>>
> > > >>>>>> What should be the use case for this? Source ports are almost
> > always
> > > >>>>> choosen randomly, so you woould get the same results as balancing
> > > >>>>> randomly.
> > > >>>>>
> > > >>>>>
> > > >>>>> Not necessarily,I came across few implementations where client
> > port is
> > > >>>> fixed (they bind() port while creating socket), but I agree that
> > most of
> > > >>>>
> > > >>>>
> > > >>> Sure.
> > > >>>
> > > >>>
> > > >>> the times source port is random.
> > > >>>
> > > >>>> The good approach would be LVS to provide options for IP+port or
> > just
> > > >>>> IP
> > > >>>> hashing.
> > > >>>>
> > > >>>>
> > > >>> But I still miss the use case. If the client always does come with
> > the
> > > >>> same port, it doesn't make a difference if the port is used too for
> > > >>> hashing
> > > >>> or just the IP.
> > > >>>
> > > >>>
> > > >> Oh..Sorry, I should have been clear.
> > > >> Here is the use case: I have a client (from out side it looks like a
> > piza
> > > >> box, but internally it has many CPU..something like blade server/ATCA)
> > > >> which initiates TCP/SCTP connections with same IP address but with
> > > >> different Port.
> > > >> So if I use SH, all these connections (potentially this client can
> > > >> initiate
> > > >> as many as 40 connections) will land on a same real server which may
> > not
> > > >> be
> > > >> what we wanted. we wanted the connections to be balanced (based on
> > > >> IP+port)
> > > >> across all the real servers. Does it make sense?
> > > >>
> > > >
> > > > Hmm, if it comes to a few thousand different ports, it would, but not
> > for
> > > > 40 (imho). ;)
> > > >
> > > > Anyway, this discussion should be held at the ML, not private. Maybe
> > > > someone else could have add some ideas.
> > > >
> > >
> > > Oh..did not notice this, this time I am including the ML group.
> >
> > Hi,
> >
> > I think that the scenario described above does make sense and that
> > none of the existing schedulers cater to it. Although persistence could be
> > used to achieve much if not all of the desired result. I would be happy
> > to consider a new scheduler that implements source address+port hashing.
> >
> > the persistence mentioned above, how does it work?Can you point to me a
> > link or tutorial on this?

The HOWTO has some information on this
http://www.austintek.com/LVS/LVS-HOWTO/HOWTO/LVS-HOWTO.persistent_connection.html

This ML archive also have various posts by myself and others on the topic.

> Also, do you have any idea if some one already working on this kind of new
> schedule?

I am not aware of anyone else doing similar work.

_______________________________________________
Please read the documentation before posting - it's available at:
http://www.linuxvirtualserver.org/

LinuxVirtualServer.org mailing list - lvs-users [at] LinuxVirtualServer
Send requests to lvs-users-request [at] LinuxVirtualServer
or go to http://lists.graemef.net/mailman/listinfo/lvs-users


krishna.sirigiri at gmail

Feb 28, 2012, 4:51 AM

Post #12 of 13 (849 views)
Permalink
Re: [lvs-users] Question on SH scheduler [In reply to]

Hi Simon,
Thanks for the link; kind of understood the concept of persistence. I
guess using this I can make sure that LVS dispatches the client
re-connection to the same real server.But to deterministically configure
connections to go to a real server we need SH.

Prasad

On Tue, Feb 28, 2012 at 3:05 PM, Simon Horman <horms [at] verge> wrote:

> On Tue, Feb 28, 2012 at 02:06:12PM +0530, krishna prasad wrote:
> > Hi Simon,
> >
> > On Tue, Feb 28, 2012 at 12:42 PM, Simon Horman <horms [at] verge>
> wrote:
> >
> > > On Tue, Feb 28, 2012 at 11:36:11AM +0530, krishna prasad wrote:
> > > > On Tue, Feb 28, 2012 at 11:18 AM, Alexander Holler <
> holler [at] ahsoftware
> > > >wrote:
> > > >
> > > > > Am 28.02.2012 06:35, schrieb krishna prasad:
> > > > >
> > > > >> On Tue, Feb 28, 2012 at 10:44 AM, Alexander Holler<
> > > holler [at] ahsoftware>*
> > > > >> *wrote:
> > > > >>
> > > > >> Am 28.02.2012 06:11, schrieb krishna prasad:
> > > > >>>
> > > > >>> On Tue, Feb 28, 2012 at 10:04 AM, Alexander Holler<
> > > holler [at] ahsoftware
> > > > >>> >*
> > > > >>>
> > > > >>>> *wrote:
> > > > >>>>
> > > > >>>>
> > > > >>>> Hello,
> > > > >>>>
> > > > >>>>>
> > > > >>>>> Am 28.02.2012 04:26, schrieb krishna prasad:
> > > > >>>>>
> > > > >>>>>
> > > > >>>>> But I strongly think that it it good to have IP+port
> hashing,
> > > for
> > > > >>>>> cases
> > > > >>>>>
> > > > >>>>> where multiple clients run on single host, in this case
> > > > >>>>>> the connections have same IP but different port. In this case
> > > also the
> > > > >>>>>> same
> > > > >>>>>> is desirable,i.e same client to the same real-server.
> > > > >>>>>> This may not make a real use case for web world, but a strong
> > > case for
> > > > >>>>>> non-web deployments like in telecom.I know LVS is increasingly
> > > used in
> > > > >>>>>> other than web services.
> > > > >>>>>>
> > > > >>>>>>
> > > > >>>>>> What should be the use case for this? Source ports are almost
> > > always
> > > > >>>>> choosen randomly, so you woould get the same results as
> balancing
> > > > >>>>> randomly.
> > > > >>>>>
> > > > >>>>>
> > > > >>>>> Not necessarily,I came across few implementations where client
> > > port is
> > > > >>>> fixed (they bind() port while creating socket), but I agree that
> > > most of
> > > > >>>>
> > > > >>>>
> > > > >>> Sure.
> > > > >>>
> > > > >>>
> > > > >>> the times source port is random.
> > > > >>>
> > > > >>>> The good approach would be LVS to provide options for IP+port
> or
> > > just
> > > > >>>> IP
> > > > >>>> hashing.
> > > > >>>>
> > > > >>>>
> > > > >>> But I still miss the use case. If the client always does come
> with
> > > the
> > > > >>> same port, it doesn't make a difference if the port is used too
> for
> > > > >>> hashing
> > > > >>> or just the IP.
> > > > >>>
> > > > >>>
> > > > >> Oh..Sorry, I should have been clear.
> > > > >> Here is the use case: I have a client (from out side it looks
> like a
> > > piza
> > > > >> box, but internally it has many CPU..something like blade
> server/ATCA)
> > > > >> which initiates TCP/SCTP connections with same IP address but with
> > > > >> different Port.
> > > > >> So if I use SH, all these connections (potentially this client can
> > > > >> initiate
> > > > >> as many as 40 connections) will land on a same real server which
> may
> > > not
> > > > >> be
> > > > >> what we wanted. we wanted the connections to be balanced (based on
> > > > >> IP+port)
> > > > >> across all the real servers. Does it make sense?
> > > > >>
> > > > >
> > > > > Hmm, if it comes to a few thousand different ports, it would, but
> not
> > > for
> > > > > 40 (imho). ;)
> > > > >
> > > > > Anyway, this discussion should be held at the ML, not private.
> Maybe
> > > > > someone else could have add some ideas.
> > > > >
> > > >
> > > > Oh..did not notice this, this time I am including the ML group.
> > >
> > > Hi,
> > >
> > > I think that the scenario described above does make sense and that
> > > none of the existing schedulers cater to it. Although persistence
> could be
> > > used to achieve much if not all of the desired result. I would be happy
> > > to consider a new scheduler that implements source address+port
> hashing.
> > >
> > > the persistence mentioned above, how does it work?Can you point to me a
> > > link or tutorial on this?
>
> The HOWTO has some information on this
>
> http://www.austintek.com/LVS/LVS-HOWTO/HOWTO/LVS-HOWTO.persistent_connection.html
>
> This ML archive also have various posts by myself and others on the topic.
>
> > Also, do you have any idea if some one already working on this kind of
> new
> > schedule?
>
> I am not aware of anyone else doing similar work.
>
> _______________________________________________
> Please read the documentation before posting - it's available at:
> http://www.linuxvirtualserver.org/
>
> LinuxVirtualServer.org mailing list - lvs-users [at] LinuxVirtualServer
> Send requests to lvs-users-request [at] LinuxVirtualServer
> or go to http://lists.graemef.net/mailman/listinfo/lvs-users
>
_______________________________________________
Please read the documentation before posting - it's available at:
http://www.linuxvirtualserver.org/

LinuxVirtualServer.org mailing list - lvs-users [at] LinuxVirtualServer
Send requests to lvs-users-request [at] LinuxVirtualServer
or go to http://lists.graemef.net/mailman/listinfo/lvs-users


horms at verge

Feb 28, 2012, 5:58 AM

Post #13 of 13 (854 views)
Permalink
Re: [lvs-users] Question on SH scheduler [In reply to]

On Tue, Feb 28, 2012 at 06:21:18PM +0530, krishna prasad wrote:
> Hi Simon,
> Thanks for the link; kind of understood the concept of persistence. I
> guess using this I can make sure that LVS dispatches the client
> re-connection to the same real server.But to deterministically configure
> connections to go to a real server we need SH.

Yes, I think that is a good description of the difference between
SH and persistence.


_______________________________________________
Please read the documentation before posting - it's available at:
http://www.linuxvirtualserver.org/

LinuxVirtualServer.org mailing list - lvs-users [at] LinuxVirtualServer
Send requests to lvs-users-request [at] LinuxVirtualServer
or go to http://lists.graemef.net/mailman/listinfo/lvs-users

Linux Virtual Server users RSS feed   Index | Next | Previous | View Threaded
 
 


Interested in having your list archived? Contact Gossamer Threads
 
  Web Applications & Managed Hosting Powered by Gossamer Threads Inc.