TSD Operational Log - Page 10
There are currently problems with the file-export feature on data.tsd.usit.no
Update:
The issue should now be resolved.
Currently, some users are not able to log in to TSD via ThinLinc. We are working on fix.
There's a license upgrade issue.
Update: the new licenses have been installed.
The machine exporting the /cluster filesystem crashed, and was down for around 10 minutes. This has caused a lot of machines mounting this share to be unresponsive.
All machines which are having trouble mounting the share will be rebooted shortly to fix any issues.
UPDATE:
All machines should now be available again. Let us know if you still have trouble accessing your machines.
Due to an issue with the virtualisation infrastructure, some VMs have been unintentionally rebooted. If you are having issues please let us know. We are investigating to fix any outstanding problems.
We need to perform a brief maintenance on the data portal, during which it will not be possible to import or export files.
We are investigating issues related to export in TSD projects
Due to a Linux operating system update some virtual machines are experiencing issues. We are working to solve the problem.
While this is ongoing, it will not be possible to reset passwords, or get new QR codes.
Since Saturday evening, there has been some instability with the Colossus file system, which affects all submit hosts and running jobs on Colossus.
Update:
The file system was up again at 09.44. All systems should work as normal, but please inform us if you have problems.
As a step in the process of getting the new storage in production, we will restrict access to Colossus from Monday March 23rd. to March 27th.
During this time, there will be no access to Colossus, and the /cluster storage, including the project areas and software modules. You can also not access these services from the TSD's virtual machines.
The HNAS areas ("durable") will not be affected by this downtime.
We will update this page with the progress of our work during the maintenance window.
Bio-IT processor software upgraded to v3.5.7.
Our TSD-users are currently experiencing issues with SPSS license. We are working on resolving this issue, but it might take a couple of working days to fix it.
UPDATE: This was solved, 2020-03-03, 13:39
VMware Horizon is currently down. This means that access to Windows virtual machines through VMware Horizon and https://view.tsd.usit.no is currently unavailable.
We are working on solving this as soon as possible.
Our apologies for the inconvenience.
--
Best regards,
TSD
The machine exporting the /cluster filesystem crashed, causing hanging mounts on machines which mounts the /cluster file system.
We're working on solving the issue.
--
Best regards,
TSD
We are working on fixing an issue affecting ssh between project VMs. While the issue persists, you may experience trouble accessing your Colossus submit host.
TSD-users cannot log in and we are investigating the cause of this and working on a fix.
We are having trouble with the Colossus NFS export, and are working to solve it.
UPDATE: Maintenance is done, and all exports of /cluster should be back to normal as of 12:58, 15-01-2020.
Due to the twice occurring crashes of the file system so far this week, we will be taking down the file system again today at 12:00, 15-01-2020 for quick maintenance.
As a result of this, /cluster will become unavailable on submit-hosts and other project VMs which mount /cluster. However, HPC-jobs running on Colossus itself will not be affected.
The machine responsible for making /cluster on Colossus available to the project machines in TSD crashed at 15:15 13-01-2020.
The services are now back up and running as expected.
For most projects this should not impact regular operations, however could create problems for projects which frequently access /cluster from their virtual machines.
We are currently checking all projects and working on getting everything back in order for the projects still affected by the outage.
--
Best regards,
TSD
SPSS is displaying a warning for license expiry. Please ignore this message. The problem will be solved soon.
We are working to solve the issue.
We had a short network-outage due to firewall-updates. The change has been reverted, and everything should be operational again.