# Flow Control

**URL:** <https://forums.percona.com/t/flow-control/3349>\
**Category:** Percona XtraDB Cluster 5.x\
**Created:** [March 19, 2014, 9:23am UTC](https://forums.percona.com/t/flow-control/3349 "2014-03-19T09:23:05Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![mmcdonough](https://avatars.discourse-cdn.com/v4/letter/m/a8b319/32.png) [@mmcdonough](https://forums.percona.com/u/mmcdonough)\
**Post date:** [March 19, 2014, 9:23am UTC](https://forums.percona.com/t/flow-control/3349/1 "2014-03-19T09:23:05Z")

</div>

I have a question about the relationship between the status variables wsrep\_flow\_control\_paused and wsrep\_flow\_control\_sent.

I have a three-node PXC 5.6 cluster (Galera 3.3 (r171)) and I am seeing numbers in the hundreds of ms for wsrep\_flow\_control\_paused, but wsrep\_flow\_control\_sent is always zero. What would be causing the values for wsrep\_flow\_control\_paused if there are no flow control events?

Here is some sample output from myq\_gadgets ‘wsrep’. The values for the other two nodes are essentially the same for p\_ms and snt. All of the values for p\_ms (wsrep\_flow\_control\_sent) are zero on all three nodes. gcs.fc\_limit is set to the default 16 so the queue on this node is exceeding the limit occasionally.

Wsrep Cluster Node Queue Ops Bytes Flow Conflct PApply Commit  
time P cnf # cmt sta Up Dn Up Dn Up Dn p\_ms snt lcf bfa dst oooe oool wind  
09:42:33 P 85 3 Sync T/T 0 29 0 5k 0 2.0M 1k 0 0 0 1 2 0 1  
09:42:38 P 85 3 Sync T/T 0 7 0 5k 0 2.0M 2k 0 0 0 1 2 0 1  
09:42:44 P 85 3 Sync T/T 0 0 0 3k 0 1.5M 426 0 0 0 1 2 0 1  
09:42:49 P 85 3 Sync T/T 0 2 0 3k 0 1.4M 648 0 0 0 1 2 0 1  
09:42:55 P 85 3 Sync T/T 0 0 0 3k 0 1.2M 570 0 0 0 1 2 0 1  
09:43:00 P 85 3 Sync T/T 0 2 0 4k 0 1.8M 1k 0 0 0 1 2 0 1  
09:43:06 P 85 3 Sync T/T 0 6 0 3k 0 1.3M 2k 0 0 0 1 2 0 1  
09:43:11 P 85 3 Sync T/T 0 28 0 5k 0 2.0M 1k 0 0 0 1 2 0 1  
09:43:17 P 85 3 Sync T/T 0 0 0 5k 0 2.0M 1k 0 0 0 1 2 0 1  
09:43:22 P 85 3 Sync T/T 0 1 0 5k 0 2.1M 968 0 0 0 1 2 0 1  
09:43:28 P 85 3 Sync T/T 0 3 0 5k 0 1.9M 226 0 0 0 1 2 0 1

---

<div class="post-metadata">

**Author:** ![mmcdonough](https://avatars.discourse-cdn.com/v4/letter/m/a8b319/32.png) [@mmcdonough](https://forums.percona.com/u/mmcdonough)\
**Post date:** [March 19, 2014, 9:26am UTC](https://forums.percona.com/t/flow-control/3349/2 "2014-03-19T09:26:13Z")

</div>

Correction. Where I said “[COLOR=#252C2F]the values for p\_ms (wsrep\_flow\_control\_sent)”, I meant “the values for snt (wsrep\_flow\_control\_sent)”.

---

<div class="post-metadata">

**Author:** ![mmcdonough](https://avatars.discourse-cdn.com/v4/letter/m/a8b319/32.png) [@mmcdonough](https://forums.percona.com/u/mmcdonough)\
**Post date:** [March 20, 2014, 11:21am UTC](https://forums.percona.com/t/flow-control/3349/3 "2014-03-20T11:21:52Z")

</div>

Apparently the problem was a bug in myq\_status. I downloaded the latest version from GitHub and I am now getting values for “snt”.
