Subscribe

How to interpret high latency and write_data on a volume?

the filer has very high avg_latency, come with very high write_latency, but read_ latency is okay. It happens at about the same time everyday, and with the same patten of similar peak value and last for same length of the time,  we know this is caused by  the backup. However, We are wondering about why write_latency is so high, Then we checked write_data graph which doesn't so obviously matching write_latency, or happened not at the same time.

1.     Does high write_latency necessarily mean that we have a lot of writes?

2.     How high write_latency is related write_data counter, As we have high write_latency,  if meanwhile a higher write_data counter should been seen as well?, and why? 

Thanks!

Re: How to interpret high latency and write_data on a volume?

Can anybody please share your ideas here?