Datastore io Delay Exceeds the Threshold

84 0 3 1

VersionAll FusionCompute versions

Issue DescriptionThe environment reports the datastore io response delay alarm. You need to find out the cause.  

Handling Process

1. The FC compares the response delay of the datastore based on the storage response delay. If the response delay exceeds the threshold, an alarm is reported. If the response delay is lower than the threshold, the alarm is automatically cleared. For different services, even if the alarm is reported, there may be no impact on the user's service when the threshold is exceeded a little.

2. Locate the drive letter based on the datastore ID in the alarm information.

Find the drive letter in the /dev/disk/by-id directory on the host.


The alarm is generated for the datastore sda10 where the VRM is located. sda10 is a partition of the sda disk. The problem is due to sda monitored by the host. If the problem is caused by other data storages, such as sdc and sdb, use the queried drive letter for further check.

3. If the alarm persists, run the iostat command on the host to check the data storage response delay.

The Await column indicates the response delay, check whether the response delay is too high. For details about the iostat command, see the following:


If the alarm is cleared, view the history log. The performance log is stored in the var/log/uvp/baklog directory of the corresponding dump log. Decompress the file named sta, and you can view the iostat.txt file. The performance log of the data storage can be viewed in the file.


As shown in the following figure:


According to the preceding analysis, the io response delay exceeds the threshold. If want to find out the root cause, it needs the analysis of the storage vendor.

  • x
  • convention:

Login and enjoy all the member benefits

Login and enjoy all the member benefits