Encountering slow physical write speeds on backup. Need to test hyperscale performance.
How can I check the physical write speed to the Gluster, StoragePool Network and the Physical Disk (Bricks) IOPS when troubleshooting performance issues?
Page 1 / 1
Hi Ron,
To test Hyperscale performance you can leverage FIO (a tool that will spawn a number of threads or processes doing a particular type of I/O action as specified by the user).
- Test Read Performance to Gluster
- groot@win-hs1 ~]# fio --name=(name of the job) --rw=(Type of I/O pattern) --direct=1 --ioengine=libaio --bs=(buffer size) --numjobs=(number of process to run) --size=(size of the file to be created) --runtime=(t --group_reporting --iodepth=(Number of I/O units to keep in flight against the file)
- Example (note - all parameters are tunable):
- For writes you should change the "--rw=" to one of the following
- --rw=read (sequential read)
- --rw=randread (random reads)
- --rw=rw (mixed sequential reads and writes)
- --rw=randrw (Mixed random reads and writes)
- # fio --name=seqread --rw=read --direct=1 --ioengine=libaio --bs=32k --numjobs=1 --size=1G --runtime=600 --group_reporting --iodepth=8
- For writes you should change the "--rw=" to one of the following
- Example (note - all parameters are tunable):
- groot@win-hs1 ~]# fio --name=(name of the job) --rw=(Type of I/O pattern) --direct=1 --ioengine=libaio --bs=(buffer size) --numjobs=(number of process to run) --size=(size of the file to be created) --runtime=(t --group_reporting --iodepth=(Number of I/O units to keep in flight against the file)
- Test Write Performance to Gluster
- oroot@win-hs1 ~]# fio --name=(name of the job) --rw=(Type of I/O pattern) --direct=1 --ioengine=libaio --bs=(buffer size) --numjobs=(number of process to run) --size=(size of the file to be created) --runtime=(t --group_reporting --iodepth=(Number of I/O units to keep in flight against the file)
- Example (note - all parameters are tunable):
- For writes you should change the "--rw=" to one of the following
- --rw=write (sequential writes)
- --rw=randwrite (random writes)
- --rw=rw (mixed sequential reads and writes)
- --rw=randrw (Mixed random reads and writes)
- # fio --name=random_read_writes --rw=randrw --direct=1 --ioengine=libaio --bs=32k --numjobs=1 --size=1G --runtime=600 --group_reporting --iodepth=8
- For writes you should change the "--rw=" to one of the following
- Example (note - all parameters are tunable):
- oroot@win-hs1 ~]# fio --name=(name of the job) --rw=(Type of I/O pattern) --direct=1 --ioengine=libaio --bs=(buffer size) --numjobs=(number of process to run) --size=(size of the file to be created) --runtime=(t --group_reporting --iodepth=(Number of I/O units to keep in flight against the file)
Since gluster writes data over the network, its also important to test network performance. You can follow these instructions for network tests.
I hope this was useful
Reply
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.