Error Description

The Linux CVM encounters memory issues, such as slow service response speed, CVM login failure, or Out of Memory (OOM).

Possible Reasons

These issues may be caused by high memory utilization of the instance, i.e., memory utilization generally stays above 90%.

Troubleshooting Approaches

1. Perform the troubleshooting procedure to check whether the memory utilization is too high.
2. See the memory issue analysis to find the causes of problems.

Troubleshooting Procedure

1. Follow the directions to check whether the memory utilization is too high.
   • If yes, proceed to the next step.
   • If not, see the memory issue analysis to find the causes of problems.
2. Log in to the CVM, run the top command, and press M to check whether there are processes in the “RES” and “SHR” columns using much memory.
   • If not, proceed to the next step.
   • If yes, perform the operations as instructed in process analysis according to the process type.
3. Run the following command to check the shared memory utilization.

   cat /proc/meminfo | grep -i shmem
   

   The following information will appear:
   
4. Run the following command to check the non-reclaimable slab memory utilization.

   cat /proc/meminfo | grep -i SUnreclaim
   

   The following information will appear:
   
5. Run the following command to check if there are huge pages.

   cat /proc/meminfo | grep -iE "HugePages_Total|Hugepagesize"
   

   The following information will appear:
   
   • If the HugePages_Total output is 0, see the memory issue analysis to find the causes of problems.
   • If the HugePages_Total output is not 0, there are huge pages. The huge page size equals to HugePages_Total * Hugepagesize. Check whether huge pages are configured by a malicious program, or if they are unnecessary, you can comment out the vm.nr_hugepage configuration item in the /etc/sysctl.conf file, and then run the sysctl -p command to abandon huge pages.

Directions

Viewing memory utilization

The free command output may vary with the Linux distributions, which is unreliable for calculating the memory utilization. Perform the following steps to view the memory utilization on the Monitoring page of the CVM console.

1. Log in to the CVM console and access the *Instances page.
2. Click the ID/Name of the instance to enter its details page. Select the Monitoring tab.
3. View memory utilization in the Memory Monitor section, as shown below:
   

Calculating memory utilization

The memory utilization is the ratio of memory used to total memory, excluding the buffer and system cache. The calculation formula is as follows:
= (Total - available)100% / Total
= (Total - (Free + Buffers + Cached + SReclaimable - Shmem)) * 100% / Total
= (Total - Free - Buffers - Cached - SReclaimable + Shmem）* 100% / Total

The required parameters Total, Free, Buffer, Cached, SReclaimable, and Shmem can be obtained in /proc/meminfo. Below is an example of /proc/meminfo.

1. [root@VM_0_113_centos test]# cat /proc/meminfo 
2. MemTotal: 16265592 kB
3. MemFree: 1880084 kB
4. ......
5. Buffers: 194384 kB
6. Cached: 13647556 kB
7. ......
8. Shmem: 7727752 kB
9. Slab: 328864 kB
10. SReclaimable: 306500 kB
11. SUnreclaim: 22364 kB
12. ......
13. HugePages_Total: 0
14. Hugepagesize: 2048 kB

The parameters are described as follows:

Memory issue analysis

If the problem persists, or an error shown below appears during your use of CVM, refer to the corresponding solutions:

• Log Error “fork: Cannot allocate memory”
• VNC Login Error “Cannot allocate memory”
• Triggering Out of Memory When There is Available Memory