This topic tells you how to install File Integrity Monitoring for VMware Tanzu (FIM).

Note When you install the FIM tile using Tanzu Operations Manager, FIM does not monitor the files on your BOSH Director. To apply FIM to the BOSH Director VM, see Installing File Integrity Monitoring on BOSH Director.

Prerequisites

Before you install FIM, you must have:

Install FIM

To install the FIM file on the Tanzu Operations Manager Installation Dashboard:

Note If you are upgrading from v1.4 or earlier, you must follow the instructions in Upgrading FIM.

  1. Download the product file from VMware Tanzu Network.

  2. Navigate to the Tanzu Operations Manager Installation Dashboard and click Import a Product to upload the product file.

  3. Under Import a Product, click + next to the version number of FIM. This adds the tile to your staging area.

  4. Click the newly added FIM tile.

Configure FIM for Linux

To configure FIM for Linux VMs:

  1. Select FIM Configuration for Ubuntu.

    FIM Configuration for Ubuntu section of the FIM configuration in Ops Manager. Click here to view a larger version of this image.

  2. Configure the following fields:

    Field Description
    Watchlist Create a list of file paths to monitor for file system events. Click Add to add file paths to the list, and click the trash can icon to remove file paths from the list.

    For more information, see Watchlist below.

    Note This field corresponds to fim.dirs in FIM v1.4 and earlier.

    Ignore patterns Create a list of files that you want FIM to ignore. Events for files matching any of the provided regular expressions are not included in the logs. Click Add to add files to the list, and click the trash can icon to remove files from the list.

    The items that you add must use Go-flavored path regular expressions. To test whether a regular expression is valid, you can use Regex101.

    For more information, see Ignore Patterns below.

    Note This field corresponds to fim.ignored_patterns in FIM v1.4 and earlier.

    Low severity tagging for frequently changed files Create a list of files to be marked as low severity. Click Add to add files to the list, and click the trash can icon to remove files from the list.

    The items that you add must use Go-flavored path regular expressions. To test whether a regular expression is valid, you can use Regex101.

    For more information, see Low Severity Events below.

    Note This field corresponds to fim.low_severity_patterns in FIM v1.4 and earlier.

    Output log format Enter a template for log entries. This template must be compatible with the golang package text/template.

    For more information about the Output log format field, see Output Log Format below.

    Note This field corresponds to fim.format in FIM v1.4 and earlier.

    Heartbeat interval (in seconds) Set the heartbeat interval as follows:
    • To activate the heartbeat interval, set the value to an integer greater than 0. If you set a negative value, an error occurs.
    • To deactivate the heartbeat interval, set the value to 0.
    The default value is 600.

    Note This field corresponds to fim.heartbeat_interval in FIM v1.4 and earlier.

    Max memory usage (in bytes) Set a limit in bytes for the maximum amount of memory, including file cache, that FIM can use per VM. The default value is 536870912 (512 MB).

    Note This field corresponds to fim.memory_limit in FIM v1.4 and earlier.

    CPU limit (percentage) Set the percentage of CPU that the FIM process can use. Integers from 1 to 100 are valid. The limit is set per core. Setting this field to 100 permits the use of one full core. The default value is 10.

    Note This field corresponds to fim.cpu_limit in FIM v1.4 and earlier.

    Enforce CPU limit Select the enforcement policy for the CPU limit (percentage):
    • Always: Ensures the CPU limit (percentage) is always enforced
    • When other processes are using CPU resources: Permits the CPU usage to exceed the limit set by CPU limit (percentage) if idle CPU cycles are available
    The default setting is When other processes are using CPU resources.

    Caution If Enforce CPU limit is set Always, verify that the CPU limit (percentage) is set high enough for FIM to execute correctly. If the limit is too strict, FIM fails to start.

    Note This field corresponds to fim.enforce_cpu_limit in FIM v1.4 and earlier.
    • Always is equivalent to fim.enforce_cpu_limit == true
    • When other processes are using CPU resources is equivalent to fim.enforce_cpu_limit == false
    Log file digest for write/create events To enable computing digests for write/create events, select Enable. When you select this option, a field for A threshold of file size beyond which digests are not calculated (in bytes) appears.

    For more information, see File Digests below.

    Note This field corresponds to fim.digests in FIM v1.4 and earlier. Setting Log file digest for write/create events to Enable is equivalent to fim.digests == [sha256].

    A threshold of file size beyond which digests are not calculated (in bytes) Enter a positive value for the threshold for the maximum size of files for FIM to hash. This field only appears if you have selected Enable for Log file digest for write/create events. The default value is 10000000.

    Note This field corresponds to fim.digest_threshold in FIM v1.4 and earlier.

    BOSH Context Adapter Enable a context adapter to recognize BOSH triggered events. If enabled, it lowers the severity of file events triggered by BOSH activities to three. The default value is Disabled.

    For more information, see BOSH Context Adaptor for File Integrity Monitoring (BCAF).
    In‑memory cache to reduce noise when files do not change To enable an in‑memory cache for security metadata, select Enable. When you select this option, a field for Maximum cache size (number of files) appears.

    For more information about the in‑memory cache, see In‑Memory Cache below.

    Note In‑memory cache to reduce noise when files do not change is currently only supported on Linux. The default value is Disabled.

    Maximum cache size (in number of files) Enter a positive integer for the maximum number of files to store in the in‑memory cache.

    This field only appears if you have selected Enable for In‑memory cache to reduce noise when files do not change. The default value is 200000, which can result in an increase in memory usage of up to roughly 100 MB.

    For more information, see In‑Memory Cache Size below.

    Important Enabling the in‑memory cache causes FIM to use additional memory. Ensure that Maximum cache size (number of files) is configured with an appropriate value using the instructions in In‑Memory Cache Size below.

    List of instance group names that will be excluded from deployment Enter a comma-separated list of instance groups that you do not want FIM deployed on.
  3. Click Save.

Configure FIM for Windows

To configure FIM for Windows VMs:

  1. Select FIM Configuration for Windows.

    FIM Configuration for Windows section of the FIM configuration in Ops Manager.

  2. Configure the following fields:

    Field Description
    Watchlist Create a list of file paths to monitor for file system events. Click Add to add file paths to the list, and click the trash can icon to remove file paths from the list.

    For more information, see Watchlist below.

    Note This field corresponds to fim.dirs in FIM v1.4 and earlier.

    Container Watchlist Create a list of file paths to monitor for file system events per container on the Windows Diego Cell. Click Add to add file paths to the list, and click the trash can icon to remove file paths from the list. For example, to monitor file system events for app files, enter C:\Users\vcap\app.

    If you do not enter a file path, FIM does not monitor any file system events for containers.

    Note The file path for generated logs from container events is relative to the file system for the Diego Cell, rather than the container.

    For example, a container event for the container file path C:\Users\vcap\app\test.html appears as a file system event in C:\proc\PID\root\Users\vcap\app\test.html, where PID is the process ID of the container.

    Ignore patterns Create a list of files that you want FIM to ignore. Click Add to add files to the list, and click the trash can icon to remove files from the list.

    The items that you add must use Go-flavored path regular expressions. When defining Ignore patterns for Windows, you must replace all single back slashes with double back slashes. To test whether a regular expression is valid, you can use Regex101. Events for files matching any of the provided regular expressions are not included in the logs.

    For more information, see Ignore Patterns below.

    Note To ignore events for files in containers, you must enter regular expressions that are relative to the file system for the Diego Cell, rather than the container. To do this, enter regular expressions that start with ^C:\proc\[^\]+\root.

    For example, to ignore all files in containers in the directory C:\Users\vcap\app, enter ^C:\proc\[^\]+\root\Users\vcap\app\.*$.

    Note This field corresponds to fim.ignored_patterns in FIM v1.4 and earlier.

    Low severity tagging for frequently changed files Create a list of files to be marked as low severity. Click Add to add files to the list, and click the trash can icon to remove files from the list.

    The items that you add must use Go-flavored path regular expressions. When defining Low severity tagging for frequently changed files for Windows, you must replace all single back slashes with double back slashes. To test whether a regular expression is valid, you can use Regex101.

    For more information, see Low Severity Events below.

    Note This field corresponds to fim.low_severity_patterns in FIM v1.4 and earlier.

    Output log format Enter a template for log entries. This template must be compatible with the golang package text/template.

    For more information, see Output Log Format below.

    Note This field corresponds to fim.format in FIM v1.4 and earlier.

    Heartbeat interval (in seconds) Set the heartbeat interval as follows:
    • To activate the heartbeat interval, set the value to an integer greater than 0. If you set a negative value, an error occurs.
    • To deactivate the heartbeat interval, set the value to 0.
    The default value is 600.

    Note This field corresponds to fim.heartbeat_interval in FIM v1.4 and earlier.

    Log file digest for write/create events Choose whether to enable computing digests for write/create events using the Enable or Disable radio buttons. If you enable digests, a field for A threshold of file size beyond which digests are not calculated (in bytes) appears after you select the option.

    For more information, see File Digests below.

    Note This field corresponds to fim.digests in FIM v1.4 and earlier. Setting Log file digest for write/create events to Enable is equivalent to fim.digests == [sha256].

    A threshold of file size beyond which digests are not calculated (in bytes) Enter a positive value for the threshold for the maximum size of files for FIM to hash. This field only appears if you have selected Enable for Log file digest for write/create events. The default value is 10000000.

    Note This field corresponds to fim.digest_threshold in FIM v1.4 and earlier.

    List of instance group names that will be excluded from deployment Enter a comma-separated list of instance groups that you do not want FIM deployed on.
  3. Click Save.

Deactivate Windows

To deactivate installing FIM on Windows VMs:

  1. In the FIM tile, select FIM Configuration for Windows.

  2. Add the instance group windows_diego_cell to the field List of instance group names that will be excluded from deployment.

  3. Click Save.

Monitor Containers with FIM

You can use FIM to monitor:

  • Garden containers on the Diego Cell VMs in VMware Tanzu Application Service for VMs (TAS for VMs)
  • Containers on the Diego Windows Cell VMs in VMware Tanzu Application Service for VMs [Windows] (TAS for VMs [Windows])

For an example log message, see Examples of Log Messages from Containers.

Monitor Garden Containers

To configure FIM to monitor Garden containers:

  1. In the FIM tile, select FIM Configuration for Ubuntu.

  2. Add the Garden container directories to the Watchlist section:

    • /var/vcap/data/grootfs/store/unprivileged/volumes/
    • /var/vcap/data/grootfs/store/privileged/volumes/

    For more information about GrootFS volumes, see Volumes.

  3. Add the following pattern to the Ignore patterns section:

    • ^/var/vcap/data/grootfs/store/(un)?privileged/volumes/[\w-]+/rootfs/.*$

    Note When files in the Garden containers are modified, changes are made to both the diff and rootfs directories. Adding this ignore pattern means that FIM ignores files and directories in the /var/vcap/data/grootfs/store/unprivileged/volumes/UUID/diff directory, where UUID is the ID of the container.

  4. Click Save.

Monitor Windows Garden Containers

To configure FIM to monitor Windows Garden containers:

  1. In the FIM tile, select FIM Configuration for Windows.

  2. Add at least one directory to the Container Watchlist section. VMware recommends that you add C:\Users\vcap\app, which is the directory for app files.

  3. Click Save.

Configure Forwarding for FIM Alerts

FIM writes all alerts to the BOSH logs for the VMs in your deployment.

  • In Linux, these logs are located in /var/log/syslog.
  • In Windows, these logs are located in C:\var\vcap\sys\log\fim-windows\filesnitch\job-service-wrapper.out.log.

You can use syslog forwarding to forward the alerts to a syslog aggregator.

  • If you are using the VMware Tanzu Application Service for VMs (TAS for VMs) tile: The syslog aggregator that you specify receives all alerts generated on TAS for VMs, including the FIM alerts. To configure system logging, follow the procedure in Configuring Logging in TAS for VMs.

  • If you are using the syslog BOSH release: You can use the syslog BOSH release to forward system logs. For more information, see syslog-release in GitHub.

When you configure syslog forwarding, ensure there is enough disk space for the logs, and that they rotate frequently. If you are not sure how often to rotate the logs, configure the rotation to occur either hourly, or when they reach a certain configured size. VMware recommends forwarding logs to a remote syslog aggregation system.

Apply Changes from Your Configuration

Your installation is not complete until you apply your configuration changes:

  1. Navigate to the Installation Dashboard in Tanzu Operations Manager.

  2. Click Review Pending Changes.

  3. Click Apply Changes to complete the FIM installation.

Verify the Installation

To verify the installation for Linux:

  1. bosh ssh into the VMs in your deployment. For more information, see BOSH SSH.

  2. Enter this command:

    touch /bin/hackertool
    
  3. Enter this command:

    grep hackertool /var/log/syslog
    
  4. Verify in the logs that a new file has been created. For example:

    CEF:0|cloud_foundry|fim|1.0.0|1|file integrity monitoring event|5| fname="/bin/hackertool" hostname="fim_1/3ad6ff1f-37e0-4b8a-80bd-d16b7f79c149" opname="CREATE" optype=1 ts=1574098829 severity=5
    


To verify the installation for Windows:

  1. bosh ssh into the VMs in your deployment. For more information, see BOSH SSH.

  2. Enter this command:

    powershell New-Item -type File /var/vcap/data/jobs/sample_file
    
  3. Enter this command:

    powershell "Get-Content C:\var\vcap\sys\log\fim-windows\filesnitch\job-service-wrapper.out.log | Select-String sample_file"
    
  4. Verify in the logs that a new file has been created. For example:

    CEF:0|cloud_foundry|fim|1.0.0|1|file integrity monitoring event|5| fname="C:\var\vcap\data\jobs\sample_file" hostname="no-job_1/ebee34c1-3300-4f5d-9557-bbef845d608c" opname="CREATE" optype=1 ts=1569953512 severity=5
    

Concepts Related to Configuring FIM

Reference the following sections when configuring FIM.

Watchlist

FIM monitors a set of critical system directories. You can configure the directories that FIM monitors by adding and removing items in the Watchlist section.

Watchlist for Linux

Below is the default list of file paths in Watchlist section of FIM Configuration for Ubuntu.

Component File Paths
System binaries and configuration /boot/grub
/root
/bin
/etc
/lib
/lib64
/opt
/sbin
/srv
/usr
/var/lib
BOSH agent /var/vcap/bosh
/var/vcap/monit/job
BOSH releases /var/vcap/data/packages
/var/vcap/data/jobs

Watchlist for Windows

Below is the default list of file paths in Watchlist section of FIM Configuration for Windows.

  • C:\Windows\System32
  • C:\Program Files
  • C:\Program Files (x86)
  • C:\var\vcap\bosh
  • C:\var\vcap\data\packages
  • C:\var\vcap\data\jobs

Ignore Patterns

Some monitored directories might contain files that you do not want FIM to monitor, such as files that change frequently. You can configure FIM to ignore these events by adding and removing items in the Ignore patterns section. Use path regular expressions.

Ignore Patterns for Linux

Below is the default list in Ignore Patterns section of FIM Configuration for Ubuntu.

Scenario List
Temporary files created when an operator or errand runs bosh ssh ^/etc/passwd.+$
^/etc/shadow.+$
^/etc/subgid.+$
^/etc/subuid.+$
^/etc/group.+$
^/etc/gshadow.+$
Temporary files created when hosts are updated ^/etc/hosts.+$
BOSH agent logs ^/var/vcap/bosh/log/.+$
Log rotation ^/var/lib/logrotate/status.*$
Monit state ^/root/\.monit\.state$

Ignore Patterns for Windows

Note There is currently no default value for Ignored patterns for Windows.

When defining Ignore patterns for Windows, you must replace all single back slashes with double back slashes. For example, to ignore all files in the directory C:\var\vcap\bosh\ignore_me\, use:

^C:\var\vcap\bosh\ignore_me\.*$

Low Severity Events

Some monitored directories might contain files that only change occasionally or files that update frequently but are low impact. You can configure FIM to log events at a lower severity by adding and removing items in the Low severity tagging for frequently changed files section. Use path regular expressions.

There are three possible severity levels:

  • 0: Used for heartbeats.
  • 3: Used for low severity events. These events are for files that match any of the provided regular expressions. This setting filters out business-as-usual events.
  • 5: Used for all other events. This is the default severity.

Low Severity Tagging for Frequently Changed Files for Linux

Below is the default list in Low severity tagging for frequently changed files section of FIM Configuration for Ubuntu.

Scenario List
When an operator or errand runs the bosh ssh a new user is created ^/etc/passwd$
^/etc/shadow$
^/etc/subgid$
^/etc/subuid$
^/etc/group$
^/etc/gshadow$
BOSH-DNS sync and new VM creation update hosts ^/etc/hosts$
Attached devices and cgroups ^/etc/mtab$
DHCP leases ^/var/lib/dhcp/dhclient.eth\d+.leases$
BOSH agent configuration changes when VM created/modified ^/var/vcap/bosh/settings.json$
BOSH agent CHMODs jobs and packages as part of bosh deployment ^/var/vcap/data/jobs$
^/var/vcap/data/packages$

Low Severity Tagging for Frequently Changed Files for Windows

Note There is currently no default value for Low severity tagging for frequently changed files for Windows.

When defining Low severity tagging for frequently changed files for Windows, you must replace all single back slashes with double back slashes. For example, to mark all files in the directory C:\var\vcap\bosh\ignore_me\ as low severity, use:

^C:\var\vcap\bosh\ignore_me\.*$

Output Log Format

By default, FIM generates messages in the Common Event Format. You can configure the output format as a Go text template using the Output log format field. For more information and examples of FIM log messages, see Log Messages.

Default Format

The default value of Output log format is:

"CEF:0|cloud_foundry|fim|1.0.0|{{.Optype}}|file integrity monitoring event|{{.Severity}}| {{.KeyValues}}"

Example output using the default Output log format configuration:

CEF:0|cloud_foundry|fim|1.0.0|0|file integrity monitoring event|0| fname="" hostname="diego_cell/8279dfa8-9f86-4bb1-8b92-65457d2ae989" opname="FILESNITCH CHECKIN" optype=0 ts=1492715822 severity=0
CEF:0|cloud_foundry|fim|1.0.0|1|file integrity monitoring event|5| fname="/etc/passwd.lock" hostname="diego_cell/8279dfa8-9f86-4bb1-8b92-65457d2ae989" opname="CREATE" optype=1 mode="-rw-r--r--" uid=0 user="root" gid=0 group="root" ts=1492715822 severity=5
CEF:0|cloud_foundry|fim|1.0.0|4|file integrity monitoring event|5| fname="/etc/passwd.17721" hostname="diego_cell/8279dfa8-9f86-4bb1-8b92-65457d2ae989" opname="REMOVE" optype=4 ts=1492715822 severity=5
CEF:0|cloud_foundry|fim|1.0.0|1|file integrity monitoring event|5| fname="/etc/group.lock" hostname="diego_cell/8279dfa8-9f86-4bb1-8b92-65457d2ae989" opname="CREATE" optype=1 mode="-rw-r--r--" uid=0 user="root" gid=0 group="root" ts=1492715822 severity=5
CEF:0|cloud_foundry|fim|1.0.0|4|file integrity monitoring event|5| fname="/etc/group.17721" hostname="diego_cell/8279dfa8-9f86-4bb1-8b92-65457d2ae989" opname="REMOVE" optype=4 ts=1492715822 severity=5
CEF:0|cloud_foundry|fim|1.0.0|1|file integrity monitoring event|5| fname="/etc/gshadow.lock" hostname="diego_cell/8279dfa8-9f86-4bb1-8b92-65457d2ae989" opname="CREATE" optype=1 mode="-rw-r--r--" uid=0 user="root" gid=0 group="root" ts=1492715822 severity=5

Note The FILESNITCH CHECKIN message is a logging marker that indicates filesnitch is operational in the absence of any file system events.

Custom Format

You can use individual fields to configure the log format. Each individual field is a named property, provided by FIM, that is replaced during the logging action.

For example:

"{{.Fname}} {{.Hostname}} {{.OpName}} {{.OpType}} {{.Metadata}} {{.Digests}} {{.Ts}}"

Example output using the above configuration:

/bin/binary plymouth CREATE 1 mode="-rw-r--r--" uid=0 user="root" gid=0 group="root" sha256=da39a3ee5e6b4b0d3255bfef95601890afd80709 1475195574

The table below lists the template values you can use:

Template Description
{{.Fname}} The name of the affected file.
{{.Hostname}} The hostname of the VM on which the file event originated.
{{.OpName}} The type of file operation in textual format. For more information about opname, see Opname and Optype below.
{{.OpType}} The type of file operation in numeric format. For more information about optype, see Opname and Optype below.
{{.Severity}} The level of importance attributed to the event. For the severity levels, see Low Severity Events above.
{{.Ts}} The point in time at which FIM received the file event in Unix epoch format.
{{.Metadata}} Key-value pairs of file mode and ownership metadata for CREATE and CHMOD events.
The information includes the file mode, owner, and group.
Metadata is only supported on Linux. For more information, see Metadata below.
{{.Digests}} Key-value pairs of hash algorithms and the hash of the modified file. For more information, see File Digests below.
{{.Json}} This string serializes an event into a standard JSON dictionary. The string is in the following format:
{"fname":"ABSOLUTE-PATH","hostname":"BOSH-VM","opname":"OPERATION-NAME","optype":OPERATION-TYPE,"metadata":{mode=FILE-MODE uid=USER-ID user="USER" gid=GROUP-ID group="GROUP"},"ts":TIMESTAMP}
For example:
{"fname":"/bin/binary","hostname":"plymouth","opname":"CREATE","optype":1,"metadata":{"mode": "-rw-r--r--","uid":"0","user":"root","gid":"0","group":"root"},"ts":1475195084}
{{.KeyValues}} This string serializes an event into a series of key-value pairs. The string is in the following format:
fname="ABSOLUTE-PATH" hostname="BOSH-VM" opname="OPERATION-NAME" optype=OPERATION-TYPE mode=FILE-MODE uid=USER-ID user="USER" gid=GROUP-ID group="GROUP" ts=TIMESTAMP
For example:
fname="/bin/binary" hostname="plymouth" opname="CREATE" optype=1 mode="-rw-r--r--" uid=0 user="root" gid=0 group="root" ts=1475195258

Opname and Optype

Opname and optype are the type of file operation in textual and numeric format, respectively. For the possible values of the two fields see the table below:

opname optype Example Linux Trigger Example Windows Trigger
FILESNITCH CHECKIN 0 This is a heartbeat message written to the log. This occurs during every Heartbeat interval.

The default interval is 600 seconds. To configure this property, see Configure FIM for Linux above.
This is a heartbeat message written to the log. This occurs during every Heartbeat interval.

The default interval is 600 seconds. To configure this property, see Configure FIM for Windows above.
CREATE 1 touch newfile.txt

echo 'content' > newfile2.txt
Powershell New-Item -type File newfile.txt

Powershell Add-Content -Path newfile.txt -Value 'content'
WRITE 2 echo 'hello world' >> file.txt Powershell Add-Content -Path newfile.txt -Value 'content'
REMOVE 4 rm file.txt Powershell rm file.txt
RENAME 8 mv file.txt file.txt.orig Powershell mv file.txt file.txt.orig
CHMOD 16 chmod 0400 file.txt Powershell icacls file.txt /grant administrators:F

Note FIM on Windows reports WRITE and CHMOD together as WRITE|CHMOD. The two operations are indistinguishable.

File Digests

FIM supports hashing monitored files on WRITE or CREATE events using the sha256 algorithm. If you enable digests, FIM includes the calculated hash for the file in the logs.

To show that content has changed, or check which version of the file is mapped to a log entry, you can calculate the sha256 value of a file and compare it to the value in the log.

Hashing is deactivated by default.

FIM sets a threshold on the size of files, in bytes, to be hashed. The default value is 10000000 bytes.

Metadata

Note Metadata is currently only supported on Linux.

FIM adds security metadata to CREATE and CHMOD file events. The table below describes the key-value pairs that the metadata uses:

Key Description
mode This key encodes Unix file system permissions and permissions behavior.

The output for setuid, getgid, and sticky bit flags differ from the standard representation in the output of the ls command.
A file with all three flags set has an output similar to ugtr-xr--r-- .
  • u indicates that setuid is set.
  • g indicates that setgid is set.
  • t indicates that sticky bit is set.
The normal output of the ls command for the same file is -r-sr-Sr-T.
uid The user ID of the owner of the file.
user The user name that corresponds to the uid. If the uid lookup fails, this is set to "unknown".
gid The group ID of the file.
group The group name that corresponds to the gid. If the gid lookup fails, this is set to "unknown".

In‑Memory Cache

Note In‑Memory Cache is currently only supported on Linux.

When you enable the in‑memory cache, the frequency of logged file events is reduced because trivial changes to files are not logged. This feature also enables you to compare the current security metadata of a file with the last known version in a CHMOD event.

Note If you enable the in‑memory cache, FIM uses additional memory. You can change the additional memory cost by configuring the cache size. See In‑Memory Cache Size below.

The table below describes the file events that are logged and ignored for CHMOD and WRITE events when the in‑memory cache is enabled:

File Event Logged Events Ignored Events
CHMOD
  • Changes to the file mode
  • Changes to user file ownership and group file ownership
  • An aggregation of file mode and file ownership changes. Only the current and last known version are logged. For example, if only the file mode changes, the metadata only includes mode and oldMode.
  • File events that are not changes to the file mode or ownership. This includes changes to file attributes that do not indicate security threats, such as timestamp changes.
WRITE
  • If Log file digest for write/create events is set to Disable, all WRITE events are logged.
  • If Log file digest for write/create events is set to Enable and the hashed SHA256 file content has not changed from the last known hash, WRITE events are not logged. This can happen if a file is overwritten with contents that are identical to the previous contents.

Note If the cache is full, because Maximum cache size (in number of files) is too small, FIM logs file events as if the cache was deactivated.

In‑Memory Cache Size

The in‑memory cache size is the maximum number of files that FIM can store at once. If there are more files being watched than can fit into the cache, the cache removes the least-recently accessed file. If the cache removes a file, the next time the file is accessed, FIM logs it as if the cache was deactivated.

On startup, FIM creates a cache consisting of all files that are descendants of the directories that are configured in Watchlist. The cache does not store full files and only stores relevant metadata about each file.

For example, if Watchlist is set to monitor the following directories:

/A
/B

And the file system has the following structure:

/
├── A
│   ├── 1
│   ├── 2
│   └── C
│       └── 3
└── B
    └── 4

Then the cache stores seven files: A, B, C, 1, 2, 3, and 4.

To determine how large to set Maximum cache size (in number of files):

  1. Set In‑memory cache to reduce noise when files do not change to Disable.

  2. Navigate to Installation Dashboard > Review Pending Changes and click Apply Changes.

  3. View the FIM stderr logs for some of your VMs. These are in /var/vcap/sys/log/fim/fim.stderr.log. The logs look similar to the following:

    ...
    2019/12/20 18:23:47 Num files added: 71011
    2019/12/20 18:23:47 Num files evicted: 0
    
  4. Record the number after Num files added for some of your VMs. This number indicates how many files would be added to the in‑memory cache if it was enabled.

  5. Set In‑memory cache to reduce noise when files do not change to Enable.

  6. Set Maximum cache size (in number of files) to 1.5 times the largest number you recorded. This accommodates for any new files that are created.

  7. Set Max memory usage (in bytes) to a value that is large enough to account for the additional memory cost for the cache. For example, 200,000 files cost approximately 100 MB in memory.

  8. Repeat steps 2 and 3.

  9. Record the number after Num files evicted. This indicates how many files could not fit into the cache. If this number is non-zero, consider increasing Maximum cache size (in number of files) to make room for these files.

check-circle-line exclamation-circle-line close-line
Scroll to top icon