<div dir="ltr"><span style="font-size:12.8px">>> Actually you would need the reduced stickiness just during the stop </span><span style="font-size:12.8px">phase - right.</span><br><div><span style="font-size:12.8px">Oh, that is good to know.</span></div><div><span style="font-size:12.8px"><br></span></div><div><span style="font-size:12.8px">While I can reduce time when waiting for </span><span style="font-size:12.8px">only "</span><span style="font-size:12.8px">stop" commands to finish, I don't think that this is worth it.</span></div><div><span style="font-size:12.8px">Because this doesn't address my problem fully.</span></div><div><span style="font-size:12.8px"><br></span></div><div><span style="font-size:12.8px">Does that mean that the reality is cruel, and there is no way to tell Pacemaker - here you have this two commands, execute them sequentially?</span></div><div><span style="font-size:12.8px"><br></span></div><div><span style="font-size:12.8px">It is all about usability for the end user.</span></div><div><span style="font-size:12.8px">As a last resort I was thinking about not providing this "do a fail-back" one-shot button to a user.</span></div><div><span style="font-size:12.8px">But instead provide "fail-back ON/OFF" switch-button, with some kind of indicator "resources are placed optimally".</span></div><div><span style="font-size:12.8px"><br></span></div><div><span style="font-size:12.8px">Anyways, maybe there </span><span style="font-size:12.8px">still </span><span style="font-size:12.8px">are some other ideas?</span></div><div><span style="font-size:12.8px">I really want to have this "one shot fail-back" rock-solid solution, and maybe I am missing here something =)</span></div><div><span style="font-size:12.8px">Or maybe it can be a feature request =)</span></div><div><span style="font-size:12.8px"><br></span></div></div><div class="gmail_extra"><br clear="all"><div><div class="gmail_signature" data-smartmail="gmail_signature"><div dir="ltr"><div><div dir="ltr">Thank you,<div>Kostia</div></div></div></div></div></div>
<br><div class="gmail_quote">On Wed, Nov 9, 2016 at 6:42 PM, Klaus Wenninger <span dir="ltr"><<a href="mailto:kwenning@redhat.com" target="_blank">kwenning@redhat.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb"><div class="h5">On 11/09/2016 05:30 PM, Kostiantyn Ponomarenko wrote:<br>
> When one problem seems to be solved, another one appears.<br>
> Now my script looks this way:<br>
><br>
> crm --wait configure rsc_defaults resource-stickiness=50<br>
> crm configure rsc_defaults resource-stickiness=150<br>
><br>
> While now I am sure that transactions caused by the first command<br>
> won't be aborted, I see another possible problem here.<br>
> With a minimum load in the cluster it took 22 sec for this script to<br>
> finish.<br>
> I see here a weakness.<br>
> If a node on which this script is called goes down for any reasons,<br>
> then "resource-stickiness" is not set back to its original value,<br>
> which is vary bad.<br>
><br>
> So, now I am thinking of how to solve this problem. I would appreciate<br>
> any thoughts about this.<br>
><br>
> Is there a way to ask Pacemaker to do these commands sequentially so<br>
> there is no need to wait in the script?<br>
> If it is possible, than I think that my concern from above goes away.<br>
><br>
> Another thing which comes to my mind - is to use time based rules.<br>
> This ways when I need to do a manual fail-back, I simply set (or<br>
> update) a time-based rule from the script.<br>
> And the rule will basically say - set "resource-stickiness" to 50<br>
> right now and expire in 10 min.<br>
> This looks good at the first glance, but there is no a reliable way to<br>
> put a minimum sufficient time for it; at least not I am aware of.<br>
> And the thing is - it is important to me that "resource-stickiness" is<br>
> set back to its original value as soon as possible.<br>
><br>
> Those are my thoughts. As I said, I appreciate any ideas here.<br>
<br>
</div></div>Have never tried --wait with crmsh but I would guess that the delay you<br>
are observing<br>
is really the time your resources are taking to stop and start somewhere<br>
else.<br>
<br>
Actually you would need the reduced stickiness just during the stop<br>
phase - right.<br>
<br>
So as there is no command like "wait till all stops are done" you could<br>
still<br>
do the 'crm_simulate -Ls' and check that it doesn't want to stop<br>
anything anymore.<br>
So you can save the time the starts would take.<br>
Unfortunately you have to repeat that and thus put additional load on<br>
pacemaker<br>
possibly slowing down things if your poll-cycle is to short.<br>
<span class=""><br>
><br>
><br>
> Thank you,<br>
> Kostia<br>
><br>
> On Tue, Nov 8, 2016 at 10:19 PM, Dejan Muhamedagic<br>
</span><div><div class="h5">> <<a href="mailto:dejanmm@fastmail.fm">dejanmm@fastmail.fm</a> <mailto:<a href="mailto:dejanmm@fastmail.fm">dejanmm@fastmail.fm</a>>> wrote:<br>
><br>
> On Tue, Nov 08, 2016 at 12:54:10PM +0100, Klaus Wenninger wrote:<br>
> > On 11/08/2016 11:40 AM, Kostiantyn Ponomarenko wrote:<br>
> > > Hi,<br>
> > ><br>
> > > I need a way to do a manual fail-back on demand.<br>
> > > To be clear, I don't want it to be ON/OFF; I want it to be<br>
> more like<br>
> > > "one shot".<br>
> > > So far I found that the most reasonable way to do it - is to set<br>
> > > "resource stickiness" to a different value, and then set it<br>
> back to<br>
> > > what it was.<br>
> > > To do that I created a simple script with two lines:<br>
> > ><br>
> > > crm configure rsc_defaults resource-stickiness=50<br>
> > > crm configure rsc_defaults resource-stickiness=150<br>
> > ><br>
> > > There are no timeouts before setting the original value back.<br>
> > > If I call this script, I get what I want - Pacemaker moves<br>
> resources<br>
> > > to their preferred locations, and "resource stickiness" is set<br>
> back to<br>
> > > its original value.<br>
> > ><br>
> > > Despite it works, I still have few concerns about this approach.<br>
> > > Will I get the same behavior under a big load with delays on<br>
> systems<br>
> > > in cluster (which is truly possible and a normal case in my<br>
> environment)?<br>
> > > How Pacemaker treats fast change of this parameter?<br>
> > > I am worried that if "resource stickiness" is set back to its<br>
> original<br>
> > > value to fast, then no fail-back will happen. Is it possible, or I<br>
> > > shouldn't worry about it?<br>
> ><br>
> > AFAIK pengine is interrupted when calculating a more complicated<br>
> transition<br>
> > and if the situation has changed a transition that is just being<br>
> executed<br>
> > is aborted if the input from pengine changed.<br>
> > So I would definitely worry!<br>
> > What you could do is to issue 'crm_simulate -Ls' in between and<br>
> grep for<br>
> > an empty transition.<br>
> > There might be more elegant ways but that should be safe.<br>
><br>
> crmsh has an option (-w) to wait for the PE to settle after<br>
> committing configuration changes.<br>
><br>
> Thanks,<br>
><br>
> Dejan<br>
> ><br>
> > > Thank you,<br>
> > > Kostia<br>
> > ><br>
> > ><br>
> > > ______________________________<wbr>_________________<br>
> > > Users mailing list: <a href="mailto:Users@clusterlabs.org">Users@clusterlabs.org</a><br>
</div></div>> <mailto:<a href="mailto:Users@clusterlabs.org">Users@clusterlabs.org</a>><br>
> > > <a href="http://clusterlabs.org/mailman/listinfo/users" rel="noreferrer" target="_blank">http://clusterlabs.org/<wbr>mailman/listinfo/users</a><br>
<span class="">> <<a href="http://clusterlabs.org/mailman/listinfo/users" rel="noreferrer" target="_blank">http://clusterlabs.org/<wbr>mailman/listinfo/users</a>><br>
> > ><br>
> > > Project Home: <a href="http://www.clusterlabs.org" rel="noreferrer" target="_blank">http://www.clusterlabs.org</a><br>
> > > Getting started:<br>
> <a href="http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf" rel="noreferrer" target="_blank">http://www.clusterlabs.org/<wbr>doc/Cluster_from_Scratch.pdf</a><br>
> <<a href="http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf" rel="noreferrer" target="_blank">http://www.clusterlabs.org/<wbr>doc/Cluster_from_Scratch.pdf</a>><br>
> > > Bugs: <a href="http://bugs.clusterlabs.org" rel="noreferrer" target="_blank">http://bugs.clusterlabs.org</a><br>
> ><br>
> ><br>
> ><br>
> > ______________________________<wbr>_________________<br>
> > Users mailing list: <a href="mailto:Users@clusterlabs.org">Users@clusterlabs.org</a><br>
</span>> <mailto:<a href="mailto:Users@clusterlabs.org">Users@clusterlabs.org</a>><br>
> > <a href="http://clusterlabs.org/mailman/listinfo/users" rel="noreferrer" target="_blank">http://clusterlabs.org/<wbr>mailman/listinfo/users</a><br>
<span class="">> <<a href="http://clusterlabs.org/mailman/listinfo/users" rel="noreferrer" target="_blank">http://clusterlabs.org/<wbr>mailman/listinfo/users</a>><br>
> ><br>
> > Project Home: <a href="http://www.clusterlabs.org" rel="noreferrer" target="_blank">http://www.clusterlabs.org</a><br>
> > Getting started:<br>
> <a href="http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf" rel="noreferrer" target="_blank">http://www.clusterlabs.org/<wbr>doc/Cluster_from_Scratch.pdf</a><br>
> <<a href="http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf" rel="noreferrer" target="_blank">http://www.clusterlabs.org/<wbr>doc/Cluster_from_Scratch.pdf</a>><br>
> > Bugs: <a href="http://bugs.clusterlabs.org" rel="noreferrer" target="_blank">http://bugs.clusterlabs.org</a><br>
><br>
> ______________________________<wbr>_________________<br>
> Users mailing list: <a href="mailto:Users@clusterlabs.org">Users@clusterlabs.org</a><br>
</span>> <mailto:<a href="mailto:Users@clusterlabs.org">Users@clusterlabs.org</a>><br>
> <a href="http://clusterlabs.org/mailman/listinfo/users" rel="noreferrer" target="_blank">http://clusterlabs.org/<wbr>mailman/listinfo/users</a><br>
<div class="HOEnZb"><div class="h5">> <<a href="http://clusterlabs.org/mailman/listinfo/users" rel="noreferrer" target="_blank">http://clusterlabs.org/<wbr>mailman/listinfo/users</a>><br>
><br>
> Project Home: <a href="http://www.clusterlabs.org" rel="noreferrer" target="_blank">http://www.clusterlabs.org</a><br>
> Getting started:<br>
> <a href="http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf" rel="noreferrer" target="_blank">http://www.clusterlabs.org/<wbr>doc/Cluster_from_Scratch.pdf</a><br>
> <<a href="http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf" rel="noreferrer" target="_blank">http://www.clusterlabs.org/<wbr>doc/Cluster_from_Scratch.pdf</a>><br>
> Bugs: <a href="http://bugs.clusterlabs.org" rel="noreferrer" target="_blank">http://bugs.clusterlabs.org</a><br>
><br>
><br>
<br>
<br>
______________________________<wbr>_________________<br>
Users mailing list: <a href="mailto:Users@clusterlabs.org">Users@clusterlabs.org</a><br>
<a href="http://clusterlabs.org/mailman/listinfo/users" rel="noreferrer" target="_blank">http://clusterlabs.org/<wbr>mailman/listinfo/users</a><br>
<br>
Project Home: <a href="http://www.clusterlabs.org" rel="noreferrer" target="_blank">http://www.clusterlabs.org</a><br>
Getting started: <a href="http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf" rel="noreferrer" target="_blank">http://www.clusterlabs.org/<wbr>doc/Cluster_from_Scratch.pdf</a><br>
Bugs: <a href="http://bugs.clusterlabs.org" rel="noreferrer" target="_blank">http://bugs.clusterlabs.org</a><br>
</div></div></blockquote></div><br></div>