Peer Global File Service Help > Advanced Topics > Conflicts, Retries, and Quarantines

Making unstructured data active at multiple locations introduces the chance of users making conflicting changes to different copies of the same file. The real-time synchronization and locking engines built into Peer Global File Service are designed to prevent these conflicts by ensuring that only one user can modify a file at a time while also making sure that all locations always have the most up to date version of a file. There are scenarios, however, where the synchronization and locking engines may not be able to prevent version conflicts. Such scenarios include network outages and file system issues.

The conflict resolution engine in Peer Global File Service is designed to handle these circumstances with a three-tiered approach backed by a combination of scans and real-time activity:

•File Conflicts – The initial state of detection of a potential version conflict. Depending on user activity, these can often be resolved automatically.

•File Retries – If certain errors are thrown when trying to synchronize a file between locations, this file will be automatically put into a retry list. Synchronization of this file will be retried every minute for a maximum of 60 attempts. The frequency of attempts and the maximum number of attempts are configurable.

•File Quarantines – These are file conflicts that could not be automatically resolved, as well as file retries that have failed after the maximum number of attempts. Files in the quarantine list will no longer be synchronized or protected with file locks until a winning file is picked through the PeerGFS user interface.

File conflicts (and potentially quarantines) can occur for any of the following reasons:

•Two users open a file at the same time or in-and-around the same time.

•A file is open at the start of a job and has been modified on a host where the configured conflict resolution strategy selects a different host as the winner.

•Two or more users have the same file open on different hosts when a collaboration job is started.

•A file was modified on two or more hosts between job restarts or network outages.

•Peer Management Center is unable to obtain a lock on a target host file for various reasons.

•Peer Management Center may conflict a file when an unexpected error occurs, or a file is in an unexpected state.

File retries can occur for any of the following reasons:

•The transfer of a file between locations is interrupted for any reason.

•The renaming of a temp file after a successful file transfer is blocked for any reason.

An example of a file conflict versus a file quarantine is as follows:

Two users have the same file open at two different locations prior to a Peer Global File Service job being enabled. When starting the job, PeerGFS will track this file as a potential conflict. If only one or no users make a change to the file, this conflict will automatically be resolved. If both users make a change, the conflict will become a quarantine.