Starting with Resilio Active Everywhere Platform v4.0.0 new type of job is available in the Management Console.
It's a dataflow driven job - File caching, provides fast access to the files on a server from multiple locations over SMB or NFS.
Configuring the job
Monitoring the cache
Peculiarities and limitations
Configuring the File caching job
File caching job is available from the Jobs tab.
Primary storage
Select the server or servers, source of the files located on a cloud or local storage. Other servers in the job will be fetching the files from the Primary storage.
Multiple servers can be the Primary storages. Note, they will also sync files with each other. Primary storages may have Read-write or Read-only permissions in the job.
MacOS Agents cannot be selected as Primary storages.
Select one or multiple primary storages to be a Priority Agent. At least one Priority Agent must be selected in the job, otherwise files won't be cleared from cache.
Caching server
Caching servers keep the cache of the files for the faster access. Files in cache are managed by the file policy.
Cache servers may have Read-write or Read-only permissions in the job. Multiple cache servers in the job also sync files with each other.
Only local server's location can be chosen as a cache path.
System requirements for cache servers
Not supported for macOS Agents, NAS devices.For Linux and Windows OS Agents check system requirements here. Additionally, for Windows OS it's required:
- Windows 10 20H2 (OS build 19042) and newer
- Windows Server 2022 (21H2) or newer with desktop experience (from release channel LTSC)
- Presence of C:\Windows\System32\cldapi.dll as a quick check for system support the functionality.
For Linux OS servers additionally "Linux cache" location need to be set. For this some pre-configuration is required. See here for more details.
"Linux": is the virtual folder path for further accessing the files over NFS.
"Linux cache": is the local cache of the above mentioned virtual folder.
Cache servers can be added to a Gateway Scale-out group. The files, requested and cached on a cache server, are automatically propagated across other cache servers in such a group. If a requested file is not available on a cache server, it's downloaded from Primary storage.
Reference server
If the job is configured to synchronise NTFS or POSIX file access permissions, Reference Agent is required in the job. Only Primary Storages can be chosen as Reference Agent.
Monitoring the Cache usage and effectiveness
Cache servers keep the files in accordance with File policy. End users may also add files to the cache storage manually over SMB or NFS.
Details about cache usage can be found in the Job runs > <selected file cache job> -> Cache server overview.
Max cache size: maximum size of the cache allowed as configured in the File policy.
Size / Number of files in cache: files that take the cache storage. This size includes:
- files fetched in the cache (pre-hydrated, pinned), either by File policy or manually by users
- files that were added manually by users, regardless of what file policy rule - Pinning or Hydration - they match, if any at all
- files in archive folder Archive folder, managed by the Agent
This size does not include:
- number of files, excluded from cache, either originated on remote Cache servers or Primary storages or added manually by users to cache folder.
- partly downloaded files in .sync directory
Size / Number of hydrated files: files that are pre-hydrated in cache by the Hydration rule in the File policy. Files that were hydrated in cache manually (i.e. not through a File policy rule) are not counted here.
Size / Number of pinned files: files that are pre-hydrated in cache by the pinning rule in the File policy. Files that were hydrated in cache manually (i.e. not through a File policy rule) are not counted here.
Percent of pinned files out of total: pinned files (by File policy) relatively the Max file cache. Files that were hydrated in cache manually (i.e. not through a File policy rule) are not counted here.
Size / Number of excluded files: files that were added to the cache locally by the users, but are excluded from cache by the Exclusion rule in the File policy or Job's Ignorelist. This parameter does not calculate the excluded files, present on a Primary storage or other Cache servers and that don't show up in cache.
Size / Number of re-downloaded files: files that were cleared from cache but then re-downloaded manually during last 7 days. Seeing large numbers here may indicate that cache file policy is misconfigured and frequently accesses flies are cleared from cache unnecessarily.
Max day of files in cache: this is the time during which a file is kept in the cache before it's cleared from it. This time is tracked by the oldest access time of the file that is cleared from cache. Pinned and excluded files are not taken into consideration when calculating this parameter.
Last time cache was filled up: days since cache was last filled up which triggered automatic clearing files from it, not necessarily it be successfull. In other words, this time does not show the date of last successful cache clearing.
Windows servers don't show the zero-sized files in file counters in the job run details, until they populated in cache by opening their location in a file browser. Until then the zero-sized files will be pending in "Files queue" for this server.
Linux servers calculate them in the cache details right away.
Peculiarities and limitations
A Priority Agent must be selected in the job, otherwise cache clearing policy won't work.
Files can be hydrated in cache not only by File policy or by a user. Third party applications may be reading or listing the files, resulting in them being hydrated in cache. Check out the Agent logs or contact support.