[Problem Description]
After the remote replication of a file system (200 GB) is configured, the actual modification amount of the file system is small, and the remote replication intermittently takes a long time to synchronize.
[Symptom Description]
The remote replication of a file system is intermittently synchronized for a long time.
[Cause]
A large number of empty files exist in the file system. However, the processing principle of empty files in remote replication is the same as that of common files. All changed data needs to be synchronized to the secondary end. For empty files, the logical size (the physical space usage is small) is copied. As a result, the synchronization time of remote replication is long. In addition, snapshots exist in the file system during remote replication synchronization. During the synchronization, the file system data changes greatly (such as deleting or modifying the file system). As a result, snapshots occupy a large amount of space. After the remote replication synchronization is complete, the storage device deletes the previously created snapshot to release the protected space.
[Location Method]
1. Check whether the version is earlier than V300R006C10.
2. Search for setHeadReqSglLis in the messager log and check the value of lba. For example, 0x1e1f38f0 indicates the location of the remote replication copy. The value is obtained by dividing the value by 2 (KB). objectId indicates the file ID and objsetid indicates the fsid of the file system.
3. Check the file path based on the fsid and file ID obtained in step 2.
4. Check ls xxx and stat xxxx at the mount point of the host based on the path obtained in step 3. If the space usage displayed by stat is less than the space usage of ls, the problem occurs.
[Emergency Measures]
Method 1: Delete the remote replication and expand the file system capacity.
Method 2: Upgrade the system to V300R003C20SPC200 and then expand the file system online.
[Solution]
Upgrade the device to V300R006C10 or a later version.
[Post-Recovery Check]
The storage system can be started and services are running properly.
[Appendix]
NA
