[prev in list] [next in list] [prev in thread] [next in thread] 

List:       keepalived-devel
Subject:    Re: [Keepalived-devel] Don't enable/disable realservers on reload
From:       J L <list1 () rrod ! net>
Date:       2010-09-07 9:41:58
Message-ID: AANLkTinBtcJuQ2vONJK6y52iBraSTOLy3DcsOhK9qYFn () mail ! gmail ! com
[Download RAW message or body]

On 30 July 2010 18:03, Vincent Bernat <bernat@luffy.cx> wrote:
> OoO Lors de la soirée naissante du vendredi 30 juillet 2010, vers 18:58,
> J L <list1@rrod.net> disait :
>
>>> But this is independant of the  "keep alive status" aspect of your code?
>>> I mean even if the real  server is deleted then recreated, its status is
>>> the one before the  reload. In this case, it would be  better to not mix
>>> patches.
>> Unfortunately, they are not independent. The clear_diff_rs function
>> decides the two RS's are not equivalent, and informs the _kernel_ to
>> remove the realserver. This results in a sub-second period where there
>> are no realservers (dammit).
>
> This  is  a  very  very  short period.   During  this  period,  existing
> connections are still working (the IPVS  table is looked at only for new
> connections).
For my purposes, why have the risk period when it isn't needed?

Regardless, the expected happened, and I have had no time to make any
improvements/changes on it. The patch is there - as I need it - I'll
leave its future up to the members of this list, as I cannot currently
justify time on anything that isn't highly urgent (sadly).

I suspect I have, however, found a bug that may be in my patch. I
haven't figured out exactly how it happens, but I sometimes get the
check being added to failed_checkers, without the RS being set to not
alive. This means the RS stays in the pool, weight never changing,
never going inactive. (If anyone has seen this bug without my patch
being involved, please let me know, so I can stop checking there)




> --
> BOFH excuse #74:
> You're out of memory
>

------------------------------------------------------------------------------
This SF.net Dev2Dev email is sponsored by:

Show off your parallel programming skills.
Enter the Intel(R) Threading Challenge 2010.
http://p.sf.net/sfu/intel-thread-sfd
_______________________________________________
Keepalived-devel mailing list
Keepalived-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/keepalived-devel

[prev in list] [next in list] [prev in thread] [next in thread]