Re: [PATCH] Document support of twos scheduler

Julian,

Many thanks for looking at my patch, and I'm sorry for any timewasting.

I don't understand why it doesn't apply since it applies cleanly for me, 
and I have even extracted
the patch from your email and that applied for me as well. Just to 
clarify, I am applying it to
commit 3c9cdaef which appears to me to be the head of the master branch at
https://git.kernel.org/pub/scm/utils/kernel/ipvsadm/ipvsadm.git. Am I 
perhaps not using the
correct git tree?

Just in case it makes any difference, and the problem is due to the 
formatting of my email,
I have attached a tar file that just contains the patch file (is that 
allowed?).

With regards,

Quentin Armitage

On 23/08/2026 13:18, Julian Anastasov wrote:
> 	Hello,
>
> On Fri, 21 Aug 2026, Quentin Armitage wrote:
>
>> The Linux kernel has supported the twos scheduler since v5.12. Make
>> sure ipvsadm users know it exists.
>>
>> Signed-off-by: Quentin Armitage <[email protected]>
>> ---
>>   SCHEDULERS | 2 +-
>>   ipvsadm.8  | 5 +++++
>>   2 files changed, 6 insertions(+), 1 deletion(-)
>>
>> diff --git a/SCHEDULERS b/SCHEDULERS
>> index 06d3357..8ddef23 100644
>> --- a/SCHEDULERS
>> +++ b/SCHEDULERS
>> @@ -1 +1 @@
>> -rr|wrr|lc|wlc|lblc|lblcr|dh|sh|sed|nq|fo|ovf|mh
>> +rr|wrr|lc|wlc|lblc|lblcr|dh|sh|sed|nq|fo|ovf|mh|twos
>> diff --git a/ipvsadm.8 b/ipvsadm.8
>> index 3d7541a..fe0f481 100644
>> --- a/ipvsadm.8
>> +++ b/ipvsadm.8
>> @@ -286,6 +286,11 @@ server as it was before.
>>   This scheduler has two flags: mh-fallback, which enables fallback to a
>>   different server if the selected server was unavailable, and mh-port,
>>   which adds the source port number to the hash computation.
> 	The patch does not apply. Probably, the old lines start
> with extra space.
>
> # patch -p1 --dry-run < /tmp/file.diff
> checking file SCHEDULERS
> checking file ipvsadm.8
> Hunk #1 FAILED at 286.
> 1 out of 1 hunk FAILED
>
>> +.sp
>> +\fBtwos\fR - random twos choice: the algorithm picks two random
>> +servers based on weights then selects the server with the least
>> +amount of connections normalized by weight. The algorithm avoids
>> +the "herd behavior" problem.
>>   .TP
>>   .B -p, --persistent [\fItimeout\fP]
>>   Specify that a virtual service is persistent. If this option is
>>
>> -- 
>> 2.55.0
> Regards
>
> --
> Julian Anastasov <[email protected]>