Splitfilesystem
- Stefan Franz
- Fabian Seifert
- Armin Franke
Splitfilesystem
Currently there are three different Versions with different features available. Splitfilesystem and SplitfilesystemV2 are in maintanace mode and can still be used in existing installations. For new installations or repositories SplitfilesystemV3 is recommended.
SplitfilesystemV3 (from tia ContentServer 3.6.0 or newer)
The files are stored within a filesystem in Base32 coded structures. The whole path is base 32 coded and divided into blocks.
The size of the blocks for folders can be choosen as 4, 8 (default) or 16. The filename is splitted with a variable number of characters, with a default of 250 chars.
The blockSize setting is stored in the filesystem in .kgs/blockSize.
Example:
FZVWO4Y\MNXW4ZTJ\M5SGC5DB\NVSXIYI\MNXW23LP\NY\MFSF643VOBYGYZLNMVXHIYLSPEXHQ3LM.InfoExample:
<repo>.contentservice.type = splitfilesystemv3
<repo>.description = Finance Repository
<repo>.saphttp.security = 1
<repo>.accessmode = rucd
<repo>.authentication.required = false
<repo>.authentication.user =
<repo>.authentication.pwdalias =
#===== FI content service configuration (filesystem|splitfilesystemv2|splitfilesystemv3|azureblobstore|s3blobstore|kgsstore)
<repo>.contentservice.splitfilesystemv3.root.1 = C:/Doc_Base
<repo>.contentservice.splitfilesystemv3.root.2 = C:/Doc_Base_Disk2
<repo>.contentservice.splitfilesystemv3.contrepinpath = true
<repo>.contentservice.splitfilesystemv3.maxfolderblocks=8
<repo>.contentservice.splitfilesystemv3.maxFileNameLength=200
<repo>.contentservice.splitfilesystemv3.rootSelectionIntervalH=24Folder separation is always / see Syntax rules for repository.cfg file
To use the splitfilesystemv3 the <repo>.contentservice.typeneeds to be splitfilesystemv3.
All parameters for splitfilesystemv3 must be configured within <repo>.contentservice.splitfilesystemv3 on a repository or template level.
Parameter | Description | Default |
|---|---|---|
root.<n> | Folder(s) where the documents and all related data are stored, it is allowed to add folder with a higher number anytime, you should not remove folder unless you do not want to access the files anymore. Using the same path more than once as root directory is not allowed. | - |
root.<n>.freespacegb | with this value it is possible to control the behavior of selecting the root folder. | -1 |
contrepinpath | create a folder in root that has the same name as the repository | false |
maxfolderblocks | create subfolder from the base32 encoded folder name with this length, this parameter can reduce the folder depth . Available values 4, 8, 16 This value cannot be changed in an existing repository it will lead to data lost and inconsistency! | 8 |
maxFileNameLength | the filename of the document encoded in base32 is maximum this long, otherwise it will be splitted in subfolders. Any number between 10 and 250 can be used. This value cannot be changed in an existing repository it will lead to data lost and inconsistency! | 250 |
rootSelectionIntervalH | Time interval to find a new root directory, when free space on disk is smaller than freespacegb. see Splitfilesystem | root folder selection rules | 12 |
root folder selection rules
Multiple root directories are allowed in SFSV3 to spread repositories across different disk/shares and avoid limit of filesystems.
Multiple roots can be configured distinguished by a order number. e.g. root.1,root.2
the algorithm tries in numbers order to pick the root
it checks the available space and compares it with the given limit. Default limit is -1.
If the algorithm finds out, that a root hasn’t enough space it tries the next.
with limit -1 a drive can be blocked. The space detection is skipped. It is then used in read only state (and / or only for adding components and informations)
if the last root has -1 it will be written until I/O-exception, means the root is “full“
the algorithm is triggered by:
if rootSelectionIntervalH is exceeded
if I/O-exception had happen while I/O-operation
Example:
<repo>.contentservice.splitfilesystemv3.root.1 = C:/Doc_Base
<repo>.contentservice.splitfilesystemv3.root.2 = C:/Doc_Base_Disk2
<repo>.contentservice.splitfilesystemv3.root.2.freespacegb=-1
<repo>.contentservice.splitfilesystemv3.root.3 = C:/Doc_Base_Disk3
<repo>.contentservice.splitfilesystemv3.root.3.freespacegb=10
<repo>.contentservice.splitfilesystemv3.root.4 = C:/Doc_Base_Disk4
<repo>.contentservice.splitfilesystemv3.root.4.freespacegb=50In this example root.1 is skiped and only used for reading because freespacegb is not set and therefore -1.
root.2 is also skipped because freespacegb is -1.
root.3 is only used when C:/Doc_Base_Disk3 has more than 10 GB available space.
root.4 is only used when C:/Doc_Base_Disk4 has more than 50 GB available space.
Is root.3 and root.4 disk space exceeded storing documents is no longer possible only reading.
Be aware that SFSV3 tries to keep documents in same folder together, when a document is stored in the original root and later a second document or component should be added in the same folder even when root.4 is used at the moment, SFSV3 will store the document in root (C:/Doc_Base). Make sure that your not used disk have some space left otherwise modification is no longer possible.
SplitfilesystemV2
The files are stored within a filesystem in Base32 coded structures. The whole path is base 32 coded and divided into blocks.
The metadata and properties of the archived document are stored in a .info-File as JSON.