<div dir="ltr"><span id="goog_1173370067"></span><span id="goog_1173370068"></span>Bartosz, if you look into what Percona guys are doing - you will see here: <a href="https://github.com/percona/percona-pacemaker-agents/blob/new_pxc_ra/agents/pxc_resource_agent#L516">https://github.com/percona/percona-pacemaker-agents/blob/new_pxc_ra/agents/pxc_resource_agent#L516</a> that they first try to use MySQL and then to get GTID from grastate.dat. Also, I am wondering if you are using cluster-wide attributes instead of node-attributes. If you use node-scoped attributes, then shadow/commit commands should not affect anything.<br>
</div><div class="gmail_extra"><br><br><div class="gmail_quote">On Mon, Jun 2, 2014 at 2:34 PM, Bogdan Dobrelya <span dir="ltr"><<a href="mailto:bdobrelia@mirantis.com" target="_blank">bdobrelia@mirantis.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="">On 05/29/2014 02:06 PM, Bartosz Kupidura wrote:<br>
> Hello,<br>
><br>
><br>
> Wiadomość napisana przez Vladimir Kuklin <<a href="mailto:vkuklin@mirantis.com">vkuklin@mirantis.com</a>> w dniu 29 maj 2014, o godz. 12:09:<br>
><br>
>> may be the problem is that you are using liftetime crm attributes instead of 'reboot' ones. shadow/commit is used by us because we need transactional behaviour in some cases. if you turn crm_shadow off, then you will experience problems with multi-state resources and location/colocation/order constraints. so we need to find a way to make commits transactional. there are two ways:<br>
>> 1) rewrite corosync providers to use crm_diff command and apply it instead of shadow commit that can swallow cluster attributes sometimes<br>
><br>
> In PoC i removed all cs_commit/cs_shadow, and looks that everything is working. But as you says, this can lead to problems with more complicated deployments.<br>
> This need to be verified.<br>
><br>
>> 2) store 'reboot' attributes instead of lifetime ones<br>
><br>
> I test with —lifetime forever and reboot. No difference for cs_commit/cs_shadow fail.<br>
><br>
> Moreover we need method to store GTID permanent (to support whole cluster reboot).<br>
<br>
</div>Please note, GTID could always be fetched from the<br>
/var/lib/mysql/grastate.dat at the galera node<br>
<div class=""><br>
> If we want to stick to cs_commit/cs_shadow, we need other method to store GTID than crm_attribute.<br>
<br>
</div>WE could use a modified ocf::pacemaker:SysInfo resource. We could put<br>
GTID there and use it the similar way as I did for fencing PoC[0] (for<br>
free space monitoring)<br>
<br>
[0]<br>
<a href="https://github.com/bogdando/fuel-library-1/blob/ha_fencing_WIP/deployment/puppet/cluster/manifests/fencing_primitives.pp#L41-L70" target="_blank">https://github.com/bogdando/fuel-library-1/blob/ha_fencing_WIP/deployment/puppet/cluster/manifests/fencing_primitives.pp#L41-L70</a><br>
<div class="HOEnZb"><div class="h5"><br>
><br>
>><br>
>><br>
>><br>
>> On Thu, May 29, 2014 at 12:42 PM, Bogdan Dobrelya <<a href="mailto:bdobrelia@mirantis.com">bdobrelia@mirantis.com</a>> wrote:<br>
>> On 05/27/14 16:44, Bartosz Kupidura wrote:<br>
>>> Hello,<br>
>>> Responses inline.<br>
>>><br>
>>><br>
>>> Wiadomość napisana przez Vladimir Kuklin <<a href="mailto:vkuklin@mirantis.com">vkuklin@mirantis.com</a>> w dniu 27 maj 2014, o godz. 15:12:<br>
>>><br>
>>>> Hi, Bartosz<br>
>>>><br>
>>>> First of all, we are using openstack-dev for such discussions.<br>
>>>><br>
>>>> Second, there is also Percona's RA for Percona XtraDB Cluster, which looks like pretty similar, although it is written in Perl. May be we could derive something useful from it.<br>
>>>><br>
>>>> Next, if you are working on this stuff, let's make it as open for the community as possible. There is a blueprint for Galera OCF script: <a href="https://blueprints.launchpad.net/fuel/+spec/reliable-galera-ocf-script" target="_blank">https://blueprints.launchpad.net/fuel/+spec/reliable-galera-ocf-script</a>. It would be awesome if you wrote down the specification and sent newer galera ocf code change request to fuel-library gerrit.<br>
>>><br>
>>> Sure, I will update this blueprint.<br>
>>> Change request in fuel-library: <a href="https://review.openstack.org/#/c/95764/" target="_blank">https://review.openstack.org/#/c/95764/</a><br>
>><br>
>> That is a really nice catch, Bartosz, thank you. I believe we should<br>
>> review the new OCF script thoroughly and consider omitting<br>
>> cs_commits/cs_shadows as well. What would be the downsides?<br>
>><br>
>>><br>
>>>><br>
>>>> Speaking of crm_attribute stuff. I am very surprised that you are saying that node attributes are altered by crm shadow commit. We are using similar approach in our scripts and have never faced this issue.<br>
>>><br>
>>> This is probably because you update crm_attribute very rarely. And with my approach GTID attribute is updated every 60s on every node (3 updates in 60s, in standard HA setup).<br>
>>><br>
>>> You can try to update any attribute in loop during deploying cluster to trigger fail with corosync diff.<br>
>><br>
>> It sounds reasonable and we should verify it.<br>
>> I've updated the statuses for related bugs and attached them to the<br>
>> aforementioned blueprint as well:<br>
>> <a href="https://bugs.launchpad.net/fuel/+bug/1283062/comments/7" target="_blank">https://bugs.launchpad.net/fuel/+bug/1283062/comments/7</a><br>
>> <a href="https://bugs.launchpad.net/fuel/+bug/1281592/comments/6" target="_blank">https://bugs.launchpad.net/fuel/+bug/1281592/comments/6</a><br>
>><br>
>><br>
>>><br>
>>>><br>
>>>> Corosync 2.x support is in our roadmap, but we are not sure that we will use Corosync 2.x earlier than 6.x release series start.<br>
>>><br>
>>> Yeah, moreover corosync CMAP is not synced between cluster nodes (or maybe im doing something wrong?). So we need other solution for this...<br>
>>><br>
>><br>
>> We should use CMAN for Corosync 1.x, perhaps.<br>
>><br>
>>>><br>
>>>><br>
>>>> On Tue, May 27, 2014 at 3:08 PM, Bartosz Kupidura <<a href="mailto:bkupidura@mirantis.com">bkupidura@mirantis.com</a>> wrote:<br>
>>>> Hello guys!<br>
>>>> I would like to start discussion on a new resource agent for galera/pacemaker.<br>
>>>><br>
>>>> Main features:<br>
>>>> * Support cluster boostrap<br>
>>>> * Support reboot any node in cluster<br>
>>>> * Support reboot whole cluster<br>
>>>> * To determine which node have latest DB version, we should use galera GTID (Global Transaction ID)<br>
>>>> * Node with latest GTID is galera PC (primary component) in case of reelection<br>
>>>> * Administrator can manually set node as PC<br>
>>>><br>
>>>> GTID:<br>
>>>> * get GTID from mysqld --wsrep-recover or SQL query 'SHOW STATUS LIKE ‚wsrep_local_state_uuid''<br>
>>>> * store GTID as crm_attribute for node (crm_attribute --node $HOSTNAME --lifetime $LIFETIME --name gtid --update $GTID)<br>
>>>> * on every monitor/stop/start action update GTID for given node<br>
>>>> * GTID can have 3 format:<br>
>>>> - XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX:123 - standard cluster-id:commit-id<br>
>>>> - XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX:-1 - standard non initialized cluster, 00000000-0000-0000-0000-000000000000:-1<br>
>>>> - XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX:INF - commit-id manually set to INF, force RA to create new cluster, with master on given node<br>
>>>><br>
>>>> Check if reelection of PC is needed:<br>
>>>> * (node is located in partition with quorum OR we have only 1 node configured in cluster) AND galera resource is not running on any node<br>
>>>> * GTID is manually set to INF on given node<br>
>>>><br>
>>>> Check if given node is PC:<br>
>>>> * have highest GTID in cluster, in case we have more than one node with „highest” GTID, we use CRC32 to choose proper PC.<br>
>>>> * GTID is manually set to INF<br>
>>>> * in case node with highest GTID will not come back after cluster reboot (for example disk failure) administrator should set GTID to INF on other node<br>
>>>><br>
>>>> I have almost ready RA: <a href="http://zynzel.spof.pl/mysql-wss" target="_blank">http://zynzel.spof.pl/mysql-wss</a><br>
>>>><br>
>>>> Tested with vanila centos galera/pacemaker/corosync - OK<br>
>>>> Tested with Fuel 4.1 - Fail<br>
>>>><br>
>>>><br>
>>>> Fuel 4.1 with that RA will not deploy correctly, because we use crm_attribute to store GTID, and in manifest we use cs_shadow/cs_commit for every pacemaker resource.<br>
>>>> This lead to cs_commit problem with different configuration in shadow copy and running configuration (running config changed by RA).<br>
>>>> "Could not commit shadow instance [..] to the CIB: Application of an update diff failed”<br>
>>>><br>
>>>> To solve this we can go in 2 ways:<br>
>>>> 1) dont use cs_commit/cs_shadow in manifests<br>
>>>> 2) store GTID in other way than crm_attribute<br>
>>>><br>
>>>> IMHO 2) is better (less invasive) and we can store GTID in corosync CMAP (<a href="http://www.polarhome.com/service/man/generic.php?qf=corosync-cmapctl" target="_blank">http://www.polarhome.com/service/man/generic.php?qf=corosync-cmapctl</a>), but this require corosync 2.X<br>
>>>><br>
>>>><br>
>>>> --<br>
>>>> Mailing list: <a href="https://launchpad.net/~fuel-dev" target="_blank">https://launchpad.net/~fuel-dev</a><br>
>>>> Post to : <a href="mailto:fuel-dev@lists.launchpad.net">fuel-dev@lists.launchpad.net</a><br>
>>>> Unsubscribe : <a href="https://launchpad.net/~fuel-dev" target="_blank">https://launchpad.net/~fuel-dev</a><br>
>>>> More help : <a href="https://help.launchpad.net/ListHelp" target="_blank">https://help.launchpad.net/ListHelp</a><br>
>>>><br>
>>>><br>
>>>><br>
>>>> --<br>
>>>> Yours Faithfully,<br>
>>>> Vladimir Kuklin,<br>
>>>> Fuel Library Tech Lead,<br>
>>>> Mirantis, Inc.<br>
>>>> +7 (495) 640-49-04<br>
>>>> +7 (926) 702-39-68<br>
>>>> Skype kuklinvv<br>
>>>> 45bk3, Vorontsovskaya Str.<br>
>>>> Moscow, Russia,<br>
>>>> <a href="http://www.mirantis.com" target="_blank">www.mirantis.com</a><br>
>>>> <a href="http://www.mirantis.ru" target="_blank">www.mirantis.ru</a><br>
>>>> <a href="mailto:vkuklin@mirantis.com">vkuklin@mirantis.com</a><br>
>>><br>
>>><br>
>><br>
>><br>
>> --<br>
>> Best regards,<br>
>> Bogdan Dobrelya,<br>
>> Skype #<a href="http://bogdando_at_yahoo.com" target="_blank">bogdando_at_yahoo.com</a><br>
>> Irc #bogdando<br>
>><br>
>><br>
>><br>
>> --<br>
>> Yours Faithfully,<br>
>> Vladimir Kuklin,<br>
>> Fuel Library Tech Lead,<br>
>> Mirantis, Inc.<br>
>> +7 (495) 640-49-04<br>
>> +7 (926) 702-39-68<br>
>> Skype kuklinvv<br>
>> 45bk3, Vorontsovskaya Str.<br>
>> Moscow, Russia,<br>
>> <a href="http://www.mirantis.com" target="_blank">www.mirantis.com</a><br>
>> <a href="http://www.mirantis.ru" target="_blank">www.mirantis.ru</a><br>
>> <a href="mailto:vkuklin@mirantis.com">vkuklin@mirantis.com</a><br>
><br>
><br>
><br>
<br>
<br>
--<br>
Best regards,<br>
Bogdan Dobrelya,<br>
Skype #<a href="http://bogdando_at_yahoo.com" target="_blank">bogdando_at_yahoo.com</a><br>
Irc #bogdando<br>
</div></div></blockquote></div><br><br clear="all"><br>-- <br><div dir="ltr">Yours Faithfully,<br>Vladimir Kuklin,<br>Fuel Library Tech Lead,<br>Mirantis, Inc.<br>+7 (495) 640-49-04<br>+7 (926) 702-39-68<br>Skype kuklinvv<br>
45bk3, Vorontsovskaya Str.<br>Moscow, Russia,<br><a href="http://www.mirantis.ru/" target="_blank">www.mirantis.com</a><br><a href="http://www.mirantis.ru/" target="_blank">www.mirantis.ru</a><br><a href="mailto:vkuklin@mirantis.com" target="_blank">vkuklin@mirantis.com</a></div>
</div>