Page History

The procedure is the same as for the Home Directory. Access is possible within the ETH network or via VPN only. You need to know the path to your share. The paths are different for each department, institute or group. If the paths are not known, they can be obtained from the respective Support group:  IT Support Group

Our NAS systems have limits on size and number of files/directories. Most operating systems don't have a special warning when the number of files/directories has reached its limit. They simply report a space problem.

Remedy can then be to clean up the data stock, or to increase the limits of the share.

The enlargements are usually made by the IT support of the department. In this case, please contact your IT support group. 

Depending on the department at ETH, the paths of the home directories point to different servers. However, as a student, doctoral student, postdoc and employee, you have a homedirectory on the servers of the ID Homedirectories with a capacity of 5 GiB.
As employees of the ID, the drives for home directories are defined by GPO and cannot be changed. But you can mount another drive at any time, which the homedirectory points to. 

With DFS :
In finder → Go  → and connect to server → in the dopdown field enter → in the dopdown field smb://d.ethz.ch/users/all/{username} → Connect 
  Example: smb://d.ethz.ch/users/all/wtester

or with a UNC path:
In finder → Go to → Connect to server Enter → in the dopdown field smb://servername/sharename{first letter of username}/{username} → Connect
  Example: smb://testnas13.nas.ethz.ch/id_service_sth_w/wtester

Advantages of using DFS:
When the ID storage services install new NAS systems, they usually get new IP addresses and DNS names. Using the DFS path eliminates the need to adjust the mount lists. The ID can adjust the reference to the new server in the DFS Target.

Advantages of the UNC path:
with slow or unstable connections a UNC path mounted home directory often behaves more error tolerant. 

If the share is already attached, you can follow this process in a terminal environment::
cd /Volumes/<share name>/.snapshots
cd \@GMT <tab> <tab> - > this shows the existing but hidden directories
cd \@GMT-2019.02.07-07.44.40/ <tab> <tab> → With Tab you can add the desired path
cd \@GMT-2019.02.07-07.44.40/ID → After that you can go to the subdirectory you are looking for

Important:
- First with cd 'only' up to the .snapshots of the share root directory.
- Afterwards with cd into a subdirectory of the desired snapshot
- After the snapshot, immediately add the /

Example:
Last login: Thu Feb 7 14:36:20 on console
id-als-docking-143:~ userid$ cd /Volumes/id_sdl_zo_data/.snapshots/
id-als-docking-143:.snapshots userid$ cd \@GMT-2019.0 <tab> <tab>
@GMT-2019.01.30-17.44.35 @GMT-2019.02.03-17.44.37 @GMT-2019.02.06-13.44.40 @GMT-2019.02.07-09.44.40
@GMT-2019.01.31-17.44.36 @GMT-2019.02.04-17.44.38 @GMT-2019.02.06-15.44.40 @GMT-2019.02.07-11.44.40
@GMT-2019.02.01-17.44.36 @GMT-2019.02.05-17.44.39 @GMT-2019.02.06-17.44.40
@GMT-2019.02.02-17.44.37 @GMT-2019.02.06-11.44.39 @GMT-2019.02.07-07.44.40
id-als-docking-143:.snapshots userid$ cd \@GMT-2019.02.07-07.44.40/ <tab> <tab>
AB/ GS/ RK/ AB_improve/ HK/ RS/ AD/ HR/ RW/
.
.

id-als-docking-143:.snapshots userid$ cd \@GMT-2019.02.07-07.44.40/AB

id-als-docking-143:ID userid$ ls
this/ and/ that/

id-als-docking-143:.snapshots userid$ cd \@GMT-2019.02.07-07.44.40/AB/this → eh voila

Open the File Explorer and navigate to the folder where a file or folder was deleted.

For your home directory, for example, enter \\d.ethz.ch\users\all\%USER% ( %USER% is replaced with your ETH username ) in the address bar.

Select the entry "Restore previous versions" in the context menu (right mouse button).

In the following dialog, select the date from which the data should be restored.

With a click on "Restore" you will restore the entire folder to the content of that date (destructiv).

• ... or you can restore individual files by double-clicking or the desired list-entry or klick Button 'open'

• select the missing files and copy them to the original location.

Our NAS systems have limits on the size and number of files/directories. Most operating systems do not have a special warning when the number of files/directories has reached its limit. They only report a space problem.

Remedy can then be found by cleaning up the database or increasing the share limits.

The extensions are usually carried out by the IT support of the department. In this case, please contact your IT support group.

• If you work with NFS clients, please consider the following points:
•  If you mount the same NFS export on a client from two different IBM Spectrum Scale NFS protocol nodes, data corruption may occur.
• The NFS protocol version that is used as default on a client operating system may differ from what you expect. If you are using a client that mounts NFSv3 by default, and you want to mount NFSv4, you must explicitly specify NFSv4 in the mount command. For more information, see the mount command for your client operating system.
•  To prevent NFS clients from having data integrity problems during failover, make sure that NFS clients are mounted with the -o hard option.
• A client must mount an NFS export using a CES IP address of a protocol node. If a host name is used, make sure that the name is unique and remains unique.
  If a DNS Round Robin (RR) record name is used to mount an NFSv3 export, data unavailability may occur due to unmounted locks. The NFS lock manager on IBM Spectrum Scale is not cluster capable.
• Clients using NFS mounts must use a timeout value for retry attempts which is slightly lower than the NFS server's grace period.
  The CES-NFS server enters the timeout period after the daemon has been restarted, or when an IP address has been shared or a new IP address has been acquired. Previously connected clients reclaim their state (e.g. - file locked, opened) within the grace period. The default value for the grace period is 60 seconds.
  The NFS client will wait for a response from the NFS server for a period of time specified by timeo before retrying requests. The timeo can be specified as an option during the connection. The value of timeo is specified in critical seconds (tenths of seconds). Clients that perform NFS connections with a retry timeout close to the NFS server's grace period may experience application failures like I/O errors.
  An example of setting the timeout value for Retry attempts to 40 seconds (overriding the Linux client's default value of 60 seconds for TCP) is - mount -o timeo=400 spectrumScaleCESIP:/path/to/exportedDirectory /localMountPoint.

*** Translated with www.DeepL.com/Translator (free version) ***

To mount a NAS share it is required to install the appropriate packages first. most of the Linux Distribtions have the package  "cifs-utils" NOT installed by default. How to install Packages is not topic here.
How ever, if you have the package installed following parameters are required as a absolute minimum:

username=[ActiveDirectory account]         <this is usually your ETH user name>

domain=D                                                  <this is the ActiveDirectory Domain name>

Beispiel: 

mount -t cifs -o username=testuser,vers=3,domain=D //nas22/beispiel_share_s1 /mnt/t4

Some OS Levels may require also the option:

Vers=2.1                                                    <this specifies explicitly the SMB Release to be used for the new session>

Please consult the man-Pages of your system for more information, avoid adding options not needed explicitly.

It is important that your system is up to date, since the SMB protocol has been enhanced havily for security reasons. Please remember that SMB1 protocol is not supported anymore! 

SMB Servers do not support multiple users on the same SMB Session. This means that if you need to connect to a share with a different user on the same NAS, this will not be allowed. You may need to umount the old share first.  

We reference here to the detailed description of the Department Biol:
  Setup a Linux system to be ready for NFS V4 and KRB5

Generally, the date of recalculation is different for each share. It depends on the creation date. The clearing period is one year.
However, we have migrated all shares existing on 7/1/2019 with this creation date to the IT Shop. So all imported shares that still exist July 1. will be billed again for another year.

However, each share owner can read the specific date in the IT Shop under 'My Products -> Storage'. The 'Valid To' indicates until when a share is paid and therefore also when the re-invoicing takes place. (see attachment)

Usually the share owners can be found in the IT support of the department / institute (ISG).

4 weeks before the re-invoicing there is a reminder to the share owner by mail.

In the ITShop the 'Valid To' field shows until when the use of the share has been paid.

Will I get a refund if I find out after 3 months that my share is too big?

• No, since the storage must also be procured and the capacity planned, the billing cycles have been set to one year.
• only in exceptional cases, which are very well justified, the complex and labor-intensive process of a refund is followed.

However, the share size can be adjusted at close to the end (directly after receipt of the mail and before the new billing for another year) or the share can be deleted without further cost consequences before billing.

At ETH, data stores are subject to a fee. Therefore, new shares can only be ordered and defined by authorized persons.
Your departmental IT support group will be happy to assist you.

The order is processed via the ITShop. Billing is done for one year in advance, analogous to software licenses.

Extensions are also invoiced directly via SAP in advance for the remaining annual term.

Each billing of a share increase must be approved by the budget manager. If the approval is denied, the corresponding resizing has to be done by the purchaser, as he is best able to assess the situation regarding possible problems and to take appropriate measures in advance. At this point, a deadline can also be agreed with the storage services.

The procedure is the same as for the Home Directory. Access is possible within the ETH network or via VPN only. You need to know the path to your share. The paths are different for each department, institute or group. If the paths are not known, they can be obtained from the respective Support group:  IT Support Group

Our NAS systems have limits on size and number of files/directories. Most operating systems don't have a special warning when the number of files/directories has reached its limit. They simply report a space problem.

Remedy can then be to clean up the data stock, or to increase the limits of the share.

The enlargements are usually made by the IT support of the department. In this case, please contact your IT support group. 

Depending on the department at ETH, the paths of the home directories point to different servers. However, as a student, doctoral student, postdoc and employee, you have a homedirectory on the servers of the ID Homedirectories with a capacity of 5 GiB.
As employees of the ID, the drives for home directories are defined by GPO and cannot be changed. But you can mount another drive at any time, which the homedirectory points to. 

With DFS :
In finder → Go  → and connect to server → in the dopdown field enter → in the dopdown field smb://d.ethz.ch/users/all/{username} → Connect 
  Example: smb://d.ethz.ch/users/all/wtester

or with a UNC path:
In finder → Go to → Connect to server Enter → in the dopdown field smb://servername/sharename{first letter of username}/{username} → Connect
  Example: smb://testnas13.nas.ethz.ch/id_service_sth_w/wtester

Advantages of using DFS:
When the ID storage services install new NAS systems, they usually get new IP addresses and DNS names. Using the DFS path eliminates the need to adjust the mount lists. The ID can adjust the reference to the new server in the DFS Target.

Advantages of the UNC path:
with slow or unstable connections a UNC path mounted home directory often behaves more error tolerant. 

If the share is already attached, you can follow this process in a terminal environment::
cd /Volumes/<share name>/.snapshots
cd \@GMT <tab> <tab> - > this shows the existing but hidden directories
cd \@GMT-2019.02.07-07.44.40/ <tab> <tab> → With Tab you can add the desired path
cd \@GMT-2019.02.07-07.44.40/ID → After that you can go to the subdirectory you are looking for

Important:
- First with cd 'only' up to the .snapshots of the share root directory.
- Afterwards with cd into a subdirectory of the desired snapshot
- After the snapshot, immediately add the /

Example:
Last login: Thu Feb 7 14:36:20 on console
id-als-docking-143:~ userid$ cd /Volumes/id_sdl_zo_data/.snapshots/
id-als-docking-143:.snapshots userid$ cd \@GMT-2019.0 <tab> <tab>
@GMT-2019.01.30-17.44.35 @GMT-2019.02.03-17.44.37 @GMT-2019.02.06-13.44.40 @GMT-2019.02.07-09.44.40
@GMT-2019.01.31-17.44.36 @GMT-2019.02.04-17.44.38 @GMT-2019.02.06-15.44.40 @GMT-2019.02.07-11.44.40
@GMT-2019.02.01-17.44.36 @GMT-2019.02.05-17.44.39 @GMT-2019.02.06-17.44.40
@GMT-2019.02.02-17.44.37 @GMT-2019.02.06-11.44.39 @GMT-2019.02.07-07.44.40
id-als-docking-143:.snapshots userid$ cd \@GMT-2019.02.07-07.44.40/ <tab> <tab>
AB/ GS/ RK/ AB_improve/ HK/ RS/ AD/ HR/ RW/
.
.

id-als-docking-143:.snapshots userid$ cd \@GMT-2019.02.07-07.44.40/AB

id-als-docking-143:ID userid$ ls
this/ and/ that/

id-als-docking-143:.snapshots userid$ cd \@GMT-2019.02.07-07.44.40/AB/this → eh voila

Open the File Explorer and navigate to the folder where a file or folder was deleted.

For your home directory, for example, enter \\d.ethz.ch\users\all\%USER% ( %USER% is replaced with your ETH username ) in the address bar.

Select the entry "Restore previous versions" in the context menu (right mouse button).

In the following dialog, select the date from which the data should be restored.

With a click on "Restore" you will restore the entire folder to the content of that date (destructiv).

• ... or you can restore individual files by double-clicking or the desired list-entry or klick Button 'open'

• select the missing files and copy them to the original location.

Our NAS systems have limits on the size and number of files/directories. Most operating systems do not have a special warning when the number of files/directories has reached its limit. They only report a space problem.

Remedy can then be found by cleaning up the database or increasing the share limits.

The extensions are usually carried out by the IT support of the department. In this case, please contact your IT support group.

• If you work with NFS clients, please consider the following points:
•  If you mount the same NFS export on a client from two different IBM Spectrum Scale NFS protocol nodes, data corruption may occur.
• The NFS protocol version that is used as default on a client operating system may differ from what you expect. If you are using a client that mounts NFSv3 by default, and you want to mount NFSv4, you must explicitly specify NFSv4 in the mount command. For more information, see the mount command for your client operating system.
•  To prevent NFS clients from having data integrity problems during failover, make sure that NFS clients are mounted with the -o hard option.
• A client must mount an NFS export using a CES IP address of a protocol node. If a host name is used, make sure that the name is unique and remains unique.
  If a DNS Round Robin (RR) record name is used to mount an NFSv3 export, data unavailability may occur due to unmounted locks. The NFS lock manager on IBM Spectrum Scale is not cluster capable.
• Clients using NFS mounts must use a timeout value for retry attempts which is slightly lower than the NFS server's grace period.
  The CES-NFS server enters the timeout period after the daemon has been restarted, or when an IP address has been shared or a new IP address has been acquired. Previously connected clients reclaim their state (e.g. - file locked, opened) within the grace period. The default value for the grace period is 60 seconds.
  The NFS client will wait for a response from the NFS server for a period of time specified by timeo before retrying requests. The timeo can be specified as an option during the connection. The value of timeo is specified in critical seconds (tenths of seconds). Clients that perform NFS connections with a retry timeout close to the NFS server's grace period may experience application failures like I/O errors.
  An example of setting the timeout value for Retry attempts to 40 seconds (overriding the Linux client's default value of 60 seconds for TCP) is - mount -o timeo=400 spectrumScaleCESIP:/path/to/exportedDirectory /localMountPoint.

*** Translated with www.DeepL.com/Translator (free version) ***

To mount a NAS share it is required to install the appropriate packages first. most of the Linux Distribtions have the package  "cifs-utils" NOT installed by default. How to install Packages is not topic here.
How ever, if you have the package installed following parameters are required as a absolute minimum:

username=[ActiveDirectory account]         <this is usually your ETH user name>

domain=D                                                  <this is the ActiveDirectory Domain name>

Beispiel: 

mount -t cifs -o username=testuser,vers=3,domain=D //nas22/beispiel_share_s1 /mnt/t4

Some OS Levels may require also the option:

Vers=2.1                                                    <this specifies explicitly the SMB Release to be used for the new session>

Please consult the man-Pages of your system for more information, avoid adding options not needed explicitly.

It is important that your system is up to date, since the SMB protocol has been enhanced havily for security reasons. Please remember that SMB1 protocol is not supported anymore! 

SMB Servers do not support multiple users on the same SMB Session. This means that if you need to connect to a share with a different user on the same NAS, this will not be allowed. You may need to umount the old share first.  

We reference here to the detailed description of the Department Biol:
  Setup a Linux system to be ready for NFS V4 and KRB5

Generally, the date of recalculation is different for each share. It depends on the creation date. The clearing period is one year.
However, we have migrated all shares existing on 7/1/2019 with this creation date to the IT Shop. So all imported shares that still exist July 1. will be billed again for another year.

However, each share owner can read the specific date in the IT Shop under 'My Products -> Storage'. The 'Valid To' indicates until when a share is paid and therefore also when the re-invoicing takes place. (see attachment)

Usually the share owners can be found in the IT support of the department / institute (ISG).

4 weeks before the re-invoicing there is a reminder to the share owner by mail.

In the ITShop the 'Valid To' field shows until when the use of the share has been paid.

Will I get a refund if I find out after 3 months that my share is too big?

• No, since the storage must also be procured and the capacity planned, the billing cycles have been set to one year.
• only in exceptional cases, which are very well justified, the complex and labor-intensive process of a refund is followed.

However, the share size can be adjusted at close to the end (directly after receipt of the mail and before the new billing for another year) or the share can be deleted without further cost consequences before billing.

At ETH, data stores are subject to a fee. Therefore, new shares can only be ordered and defined by authorized persons.
Your departmental IT support group will be happy to assist you.

The order is processed via the ITShop. Billing is done for one year in advance, analogous to software licenses.

Extensions are also invoiced directly via SAP in advance for the remaining annual term.

Each billing of a share increase must be approved by the budget manager. If the approval is denied, the corresponding resizing has to be done by the purchaser, as he is best able to assess the situation regarding possible problems and to take appropriate measures in advance. At this point, a deadline can also be agreed with the storage services.