Closed
Bug 663945
Opened 13 years ago
Closed 13 years ago
decommission EQL breakpad-dumps volume and mrdb70 server
Categories
(mozilla.org Graveyard :: Server Operations: Projects, task)
mozilla.org Graveyard
Server Operations: Projects
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: nmaul, Assigned: jabba)
Details
We would like to shut off this iSCSI volume and server. The volume appears to be mounted only by nm-sun-xf01, and sees extremely infrequent access (most recently, a very small hit on April 7 of this year). I have been told it contains some data that needs to be imported into hbase before we can get rid of it. Additionally, the server mrdb70 is reportedly related to this, but I cannot log in to confirm. I do know that it is not directly connected to the iSCSI mount (although it could be, through an NFS relay on nm-sun-xf01).
Comment 1•13 years ago
|
||
Laura, The easiest way I can think of to import the data would be to rely on the existing python code. i.e. write a script that traverses the data and then makes the appropriate call in hbaseClient.py to save the items.
Assignee | ||
Comment 2•13 years ago
|
||
mrdb70, I was told, used to be some ancient production postgres server for socorro and possibly contains the data associated with the raw dumps that are living on that iSCSI volume. I was told by aravind before he left, that lars has some knowledge about this stuff and that currently we have just been sitting on it pending some sort of data retention policy decision. I'd rather just dump the crashes into hbase, archive the database in some way, if desired and kill of both the server and the iscsi volume.
Comment 3•13 years ago
|
||
Hrm. okay, I was thinking the data was along the lines of the NFS directory structure. If it is a db dump, then the easiest way we could deal with the data would be to just tar it up and store it directly in HDFS in the Socorro PHX cluster. It can sit there for as long as it needs to as an opaque blob.
Assignee | ||
Comment 4•13 years ago
|
||
I think there is that too. I believe there are raw dumps sitting in NFS on that iscsi volume and then in addition to that, there is the postgres database, which could also be dumped, I suppose. Need mpressman and lars to look through and figure out what we are dealing with here...
Updated•13 years ago
|
Assignee: server-ops → jdow
Comment 5•13 years ago
|
||
mpressman is out sick. can someone get me access to the volume in question? I can prepare a list of what's the files there are.
Assignee | ||
Updated•13 years ago
|
Component: Server Operations → Server Operations: Projects
Comment 6•13 years ago
|
||
This data is two years old. Delete it. a=laura
Assignee | ||
Comment 7•13 years ago
|
||
mrdb70 has been shutdown, marked as spare in inventory and the iscsi volumes have been deleted.
Status: NEW → RESOLVED
Closed: 13 years ago
Resolution: --- → FIXED
Comment 8•13 years ago
|
||
thank you, the knowledge of the existence of this data has been haunting me for two years. I can finally let it go.
Updated•9 years ago
|
Product: mozilla.org → mozilla.org Graveyard
You need to log in
before you can comment on or make changes to this bug.
Description
•