Summary

• Page properties (1 modified, 0 added, 0 removed)

Details

Page properties

Content

...	...	@@ -1,5 +1,20 @@
1		-=== **Collaboratory Drive issue (2023-04-04)** ===
	1	+(% class="wikigeneratedid" %)
	2	+=== **Collaboratory Dive and Lab work (2023-04-09)** ===
2	2
	4	+We are announcing a down time for the Collaboratory Drive this coming Sunday April 9 from 8 AM CET lasting potentially all day. This will allow us to activate a secondary Drive server, mirror copy of the primary server, which we intend to use for backups only.
	5	+
	6	+During the coming days, we will also be tentatively deploying and exercising new policies in the Drive.
	7	+
	8	+1. We will limit or disable the creation of [[core dumps>>https://en.wikipedia.org/wiki/Core_dump]] in the Drive by the Lab. An important number of large core dumps have recently been filling up the Drive, which has caused the disk full situation earlier this week much faster than would have happened otherwise.
	9	+1. We will limit the maximum file size that can be generated by the Lab, both in the Drive and in the local file system of the user's Lab container. This file size will be aligned with the maximum upload size limit to the Drive which is currently 1 GB. The reasons for this are:
	10	+11. In the Drive: the alignment of the policy of upload of files and Lab creation of files. The Drive is not intended for very large files. Collabs have a Bucket storage which is optimized for this purpose.
	11	+11. In the local file storage of Lab containers: each Lab container is only accessible to a single user, but the storage resource is shared at the OpenShift level by all the pods running on a same worker node. A Lab container therefore has the possibility of causing a disk full condition on the other users' containers. The file size limit will reduce this risk. Users are reminded that stress testing any resources provided on the EBRAINS RI is strictly forbidden without the explicit authorization from EBRAINS Technical Coordination.
	12	+1. We will "move" (copy and delete) files larger than the 1 GB limit from the Drive to the Bucket of the same collab. Files that are "moved" will be deleted from the Drive and a file with the same name extended with "_moved.txt" will be put in its place. This file will inform users to check the Bucket for their file. Example: the file **//MyTalk.mp4//** will be deleted and replaced by a file //**MyTalk.mp4_moved.txt**.//
	13	+
	14	+The new policies are not technically releases but they will be announced in The Collaboratory [[Releases>>doc:Collabs.the-collaboratory.Releases.WebHome]] page.
	15	+
	16	+=== **Collaboratory Drive issue (2023-04-04) (Resolved)** ===
	17	+
3	3	Some users are still experiencing issues while trying to access the Drive. We are investigating the issue and will update this announcement as soon as we have more information.
4	4
5	5	The issues on the drive have been resolved.
...	...	@@ -6,6 +6,8 @@
6	6
7	7	We are still looking to recover files that were lost due to the disk full condition earlier this week. It appears that there were no files lost that were older than Thursday March 30. If you identify any such file, please contact EBRAINS Support.
8	8
	24	+Update: We have looked into all our options for recovering lost data from the event earlier this week. We're sorry to have to announce that no more data will be recoverable. We are however going to reattempt a continuous backup strategy to limit the risk and potential extent of data loss.
	25	+
9	9	=== **Collaboratory Drive repair (2023-04-03) (Resolved)** ===
10	10
11	11	(% class="wikigeneratedid" %)