Tomorrow, December 15, we will be physically re-locating the servers that are serving your OPI’s with name resolution and backup. Servers will be taken off-line at approximately 10.00 (CET).
This means that:
- Any backups will be unavailable during most of the day.
- Backups may be interrupted, a long backup that is still running late tomorrow morning (CET) will fail.
- Possibly, but not likely, backups running on Friday may be affected.
- OPI will not be able to authenticate to our servers during this time. Due to this, DNS servers will not be accepting updates to IP-numbers, and new certificates will not be generated.
Our plan is that everything should be up and running again by the end of the day. If you notice problems with your OPI, please have some patience and if you still have problems on Friday morning, please restart your OPI.
The servers will be moved from Sjöbo and Ystad which is the current location to Landskrona and Lund, so your data will still never leave Sweden.
/PA & Tor
Servers are now back up and running again, the DNS is updated and we have successfully been running a couple of backups using the new location.
For those of you using our backup systems, we recommend that you reboot your OPI in order to synchronize the backup scripts with the new location.
/PA
During Dec 23’rd we experienced a 5min downtime on what seemed to be a DNS lookup failure on one of our servers.
This caused our servers to reply with a timeout response that the backup-client could not handle and the backup process of the OPI stalled.
If you have received a message in the admin-email of OPI with the content
“start: Job is already running: anacron”, this is most likely what has occured for you.
The solution to this is to login to OPI and restart the unit. This will trigger a filesystem check of the backup and then it will be remounted correctly. No data is lost in the process, and you shoule be able to access your backups as usual after the reboot. Running the filesystem check can take a while though if you have a lot of data, so please have some patience.
/PA
We have had a few more reported issues with connections failures to the new server. We have adjusted some parameters, and I hope this will solve the issue.
The problem for is us that it is very hard for us to trigger this, it is a matter of more simultaneous connections to the backup than what we currently can trigger in our test systems.
/PA