R1Soft backup may fail in some circumstances, even though no change occurred, the agent and kernel hcp module properly installed with message that the replication driver was not detected as shown below :
8/26/20 5:14:58 PM Error Protected Machine An exception occurred during the request. The replication driver was not detected (detail: An error occurred while loading the hcpdriver module, please check your system logs; Please run "r1soft-setup --get-module " on your Agent to install one. ). 8/26/20 5:14:58 PM Error Manager Error replicating device / (99b71fd6-4676-4fdf-b6fe-8923a64a0531) 8/26/20 5:14:58 PM Error Manager Agent reported error during requested operation
Looking at the protected machine system logs, we can observe a “call trace”, memory related entries and the following regarding the “replication driver” (hcp) :
hcp: ERROR: Could not init mrf hash!
Knowing that the module is already installed and the agent running, we can see that the module is not loaded in the kernel by using the following command :
lsmod | grep hcpdriver
If the above did not returned anything (like in this case), it is usually caused by memory fragmentation.
The recommended fix would be rebooting the server if possible, or by compacting the memory and clearing the disk cache. The second option is only if a reboot is unthinkable, as it may causes issues in some circumstances. I would definitely avoid and especially on database servers.
Here is the procedure for the second approach :
1. Compact the memory :
sync; echo 1 > /proc/sys/vm/compact_memory
2. Clear the disk cache :
sync; echo 1 > /proc/sys/vm/drop_caches
Use the “number to echo” to drop_caches as followed, by starting by “1” and see between each steps if it solved your issue :
echo 1: Clears PageCache only
echo 2: Clears dentries and inodes
echo 3: Clears PageCache, dentries and inodes
WARNING : Use option 3 for last resort only. You should definitely consider a reboot since serious impact may occur in production.
3. Restart the R1Soft agent :
/etc/init.d/cdp-agent restart