about SNAPMIRROR Splitter critical threshold breached

[SimsonChan]

"HA Group Notification from node1 (SNAPMIRROR SYNCHORONOUS EXCEPTION: Splitter critical threshold breached [OWM]) NOTICE" alert we received recently. However, we don't find any error on the snapmirror sync. All synced normally. 

I tried to find more information about how this Splitter work in snapmirror. but seems fewer information for public.

I can only found "The splitter sends the I/O to both primary and secondary storage systems over the interconnect replication network." on the site. 

My question is 

1.       How does the splitter work in snapmirror?

2.       In what circumstances that splitter threshold will be breached? 

3.       What is the splitter threshold? where I can print the policy out? Seems it can be turn on/off only?

4.      How to troubleshoot if we encounter such error for splitter?

[hmoubara]

Hello,

The splitter is an option (are-data-ops-sequentially-split) that you can either  enable/disable under snapmirorr policy.

This parameter specifies whether I/O, such as write, copy-offload and punch-holes, are split sequentially, rather than being run in parallel on the source and destination. Spliiting the I/O sequentially will make the system more robust, and less prone to I/O errors. However, it will also make I/O performance slower.

I have pasted the below link for further information:

https://docs.netapp.com/ontap-9/topic/com.netapp.doc.dot-cm-cmpr-960/snapmirror__policy__create.html

Hope this answer your questions and concern.

Thanks 

[SimsonChan]

Hi hmoubara, thanks for your answering. 

I know it can be enable/disable by command snapmirror policy. But what behavior would causes breached Splitter threshold? And what's the definitions. I'm looking for the detail information about. eg. the data ops didn't split and sent? Thanks.

[paul_stejskal]

Do you have a case open? I'd wanna review deeper if we had a case open to answer that question. We don't have much on that message unfortunately.

I would say that if you are not having issues, it shouldn't be too big of an issue. But I think we may need to dig deeper.

[SimsonChan · Accepted solution]

According to NetApp consultant's reply.

It just recorded the event that one of the snapmirror synchronous resource pool (OWM) has reached a high threshold, due to which a recycling was activated to reclaim the unused ones. This is indicative of a workload operating on many small files. That's what triggered this event.