This is an old revision of the document!
The data01
storage server (22 TB) and its backup are at their end of life and a new environment (83TB) is available. We will call it and its backup device data02
for now. Next to the data01
server there was a bulk storage server data03
(34TB used) that was not backed up. Both the old data01
and data03
will be merged now to data02
.
The new data02
is configured identically to the old data01
with the same PI group structure and access rights. The current PI structure :
Name | Name |
---|---|
pi-aarts | pi-aartsma |
pi-allan | pi-bouwmeester |
pi-dedood | pi-frenken |
pi-gast | pi-groenen |
pi-heinrich | pi-huber |
pi-kraft | pi-oosterkamp |
pi-orrit | pi-schmidt |
pi-semrau | pi-vandermolen |
pi-vanexter | pi-vanhecke |
pi-vannoort | pi-vanruitenbeek |
The Storage server consist out of two elements, a server machine (Dell R630) that handles the network access, the file I/O, user and access rights and connects via fibre channel to an extension box (PowerFault MD36xx series). Between the server and the extension box a redundant pair of fibre channels is setup, with two independent interface cards that distribute the workload (load balancing). The system disk of the servers is configure RAID1 for High Availability.
The hardware is bought for expansion. Currently the disk extension box acquired contains 20 disks, and can easily be extended with additional disks to a total of 60 disks. So the current 83TB can be extended to about 250TB by simply acquiring more disks. When the disk cabinet is filled an additional cabinet could store an additinal; 250TB of space without any major change to the hardware configuration. Extending the partition size of the shares is done dynamically, so with the addition of new physical disks you will only see that the total size of free space increases. You do not have to use a different share mapping.
The way the data is stored across the disk is in the form of Pools. This means that data is written in a redundant format across all disks simultaneously increasing write performance while making it resistant to multi disk failure. The system is configured such that it will report disk failures so system managemen can take appropriate actions. Currently, the maintenance contract is for four years, while disk replacement can be performed more or less indefinitely.
In order to save your data onto you own PI area one needs to mount the appropriate share. Above you see the avaialble share names. You only have access to the PI share you have been assigned to by system management. If you need access to other PI share, please contact the PI and system amangement.
There are two ways to connect to the shares. The simple way is from the meetnet
PC or your personal desktop. Alternatively you cann connect to these shares using an ass tunnel. Both methods are described below.