The IT
Incidents and Maintenance
Maintenance SimLab 20. Feb. 2019, 15:00
Our git-repository-server (simlab.ww.uni-erlangen.de) will be offline during a bigger update and maintenance period on Wednesday the 20. Feb. 2019.
Data synchronized during the update process could be lost.
Maintenance successful - ηὕρηκα
- New certificate employed. Now external use will be much easier.
- Database updated.
HPC (Cluster)
New Simulation Nodes
| hostname | access | Hardware | \~ | \~ | Status |
|---|---|---|---|---|---|
| sim01 | cl2 | 12 cores, \~24 GB RAM | \~ | \~ | running |
| sim02 | cl2 | 16 cores, \~48 GB RAM | \~ | \~ | running |
| sim03 | cl2 | 18 cores, \_128 GB RAM | \~ | \~ | running |
| sim04 | cl1 | 12 cores, \~96 GB RAM | Fedora 34 | \~ | testing |
| sim05 | cl2 | 48 cores, \_168 GB RAM | \~ | \~ | running |
| sim06 | special | 6 cores, \~\_8 GB RAM | \~ | \~ | limited |
| sim08 | special | 8 cores, \~16 GB RAM | \~ | \~ | limited |
historic machines
For running old simulations we keep some old systems. They can be activated instead of the current simulation nodes.
| sim32 | cl1 | \~8 cores, \~4 GB RAM | Fedora 32 (2021) | \~ | maintenance |
|---|---|---|---|---|---|
| sim30 | cl1 | \~8 cores, \~4 GB RAM | Fedora 30 (2020) | \~ | on demand |
| sim29 | cl1 | \~6 cores, \~4 GB RAM | Fedora 29 (2019) | \~ | on demand |
| sim27 | cl1 | \~6 cores, \~4 GB RAM | Fedora 27 (2018) | \~ | on demand |
| sim25 | cl1 | \~6 cores, \~4 GB RAM | Fedora 25 (2017, lammps with atomistica) | \~ | on demand |
| sim23 | cl1 | \~4 cores, \~4 GB RAM | Fedora 23 (2016) | \~ | on demand |
| sim1204 | cl1 | \~4 cores, \~4 GB RAM | Ubuntu 12.04 (2014) | \~ | on demand |
If you have demand for some special configurations, more cores more memory on one of the simulation nodes, or you want to try some historic machines just ask the IT to reconfigure the systems
Intel ME Security Vulnerability (INTEL-SA-00086)
12.05.2017 - 16:00
The security vulnerability public since 1. Mai 2017 (INTEL-SA-00075, INTEL-SA-00086) has now been categories by Intel as severe.
Most PCs, Laptops and Servers based on Intel CPUs can be captured, crashed or damaged, remote or with an USB-stick, powered up or not.
Intel published a detection tool for testing the own system:
Downtime of HPC clusters LiMa, TinyFAT & parts of Woody on Mon, May 15.
12.05.2017 - 16:00
Do to urgent work on the power grid, the HPC clusters LiMa, TinyFAT & parts of Woody (w10xx = :sb) have to shut down on Monday, May 15th starting at 7 o'clock in the morning. As usual, jobs that would collide with the downtime will be postponed.
Check the MOTD when logging into the clusters for updates.
New Repositories: SVN ⟶ GIT
28.04.2017 - 14:00
Starting from today our old subversion repositories on ww8depot.ww.uni-erlangen.de/svn/ will be moved to our new
SimLab(git) repository management system, which is based on GIT.
In this process the repositories will be converted to GIT.
To access the repositories on SimLab you have to log in there once and ask an administrator (me) to give you rights for the repository. This has only to be done for the first repository, all other access rights will be added automatically by the system (hopefully).
If you can not access your repositories any more or have questions about the usage of GIT.
If the demand for counselling is big enough and questions arise about how to plan and handle a project with the new GIT-workflow we should hold a panel and discuss best practices.
<< Newer entries | Older entries >>
The IT is back
30.09.2016 - 14:00
All systems working again, the Cloud services included.
System Emergency shutdown - back to normal
23.09.2016 - 16:30
The new identity management system (IDM) is working like a charm.
- All clients in the institute are back to normal.
- Most of the server services are working again, network shares and SVN repositories.
But one last thing missing are the cloud systems with calendars and filesharing. You can use the ww8Cloud again and access your files and personal calendar, but sharing and the institut calendars are still missing.
System Emergency shutdown - most systems back
22.09.2016 - 19:08
We lost our identity management system (IDM) today which is managing all our clients and user accounts.
A new IDM has been prepared and tested during the last months and has taken over now.
The bad news is all user accounts got lost and have to be recreated (except the ones of people with test accounts in the new systems). Everyone has to get a new password. I will send new temporary passwords by email if the email-address is available to me, unfortunately the email-addresses of all our users had been saved in the broken IDM.
All the clients in the institute (except for ww8-node1 and ww8-2029-1) are back to normal, home directories are available again.
Sometimes there is strange behavior when sshing around. Workaround is to kinit and source the ~/.bashrc . This will be solved tomorrow.
The SVN repositories are still not available, but will be back tomorrow too.
The Ownclouds (owncloud and fass-cloud) are not working, the need a little bit more tuning, hopefully they will be back tomorrow afternoon too.
The student laptops which have been borrowed by some students (Bergmann and Monnavari) will loose the ability to login during the next weeks and have to be returned to the institute to be configured for the new IDM.
System Emergency Shutdown
22.09.2016 - 11:19
All our systems will be shut down for at least one our. After this all passwords have to be reset.
- Home directories and Login to clients in the institute will not be available.
- SVNs will not be available.
- Owncloud and other services will not be available, maybe for an even longer time.
- The cause of all these troubles is our Certificate system which died yesterday afternoon. Certificates can not be renewed because of some internal problems.
The fastes way to solve the problem is to just set up a new one.
