Difference between revisions of "Notices"
Line 46: Line 46:
  
 
::'''UPDATE:  1/29/20'''  We are applying alternative pre-bug kernels to those  Linux clients that are available to be rebooted (to activate these kernels.)  As more become available, we'll reboot them as well.  We're hoping these alternative kernels will prevent the hangups until officially patched kernels become available.
 
::'''UPDATE:  1/29/20'''  We are applying alternative pre-bug kernels to those  Linux clients that are available to be rebooted (to activate these kernels.)  As more become available, we'll reboot them as well.  We're hoping these alternative kernels will prevent the hangups until officially patched kernels become available.
−
 
−
 
−
<hr>
 
−
:'''WHAT:'''  Slow logins to gradx.cs.jhu.edu and gradz.cs.jhu.edu
 
−
 
−
:'''WHEN:'''  First reported 1/28/20
 
−
 
−
:'''DETAILS:'''
 
−
 
−
::Users might be noticing very slow logins to gradx and gradz.  It may take a couple of minutes for the login process to complete.  We are actively searching for a solution to this problem, which currently, is not affecting our other systems.  For a complete list of our available systems, please see: [https://support.cs.jhu.edu/wiki/Category:Linux_Clients https://support.cs.jhu.edu/wiki/Category:Linux_Clients]
 
−
<hr>
 
−
<br>
 
  
 
=== PLANNED CS SERVICE OUTAGES ===
 
=== PLANNED CS SERVICE OUTAGES ===

Revision as of 19:49, 31 January 2020


SERVICE OUTAGES

UNPLANNED CS SERVICE OUTAGES

WHAT: Linux kernel bug causing some of our Fedora 31 lab systems to hang.
WHEN: First identified in CS, last week of January, 2020. (See UPDATE: below the DETAILS: section.)
DETAILS:
As you know, recently, we upgraded our lab Linux clients to Fedora 31. Over the past several days, we've discovered some odd behavior from some of the lab systems. Upon further investigation, we have learned that there is a kernel bug that is causing some systems to hang (requiring a reboot) when certain NFS operations are performed. (NFS is the system that allows our Linux clients to access the home directories, certain apps, other shared file systems.)
We're seeing that the bug is documented and that it affects not only Fedora 31, but some prior versions as well. It just happens to be that we are seeing its effects now.
This is a Linux kernel issue and there are Linux folk working on trying to get this bug patched, but we do not know how long it will take.
In the meantime, CS IT is looking into work-arounds to bypass this bug as much as possible.
There is no way to predict if or when this bug will cause a system to hang. Some of our systems may never notice it. We think it might happen to some systems that are heavily utilizing NFS operations, but again, we're not sure.
Our recommendation is that when you're editing files, etc. save your work on a regular basis, just in case. This is a good practice to follow anyway.
IMPORTANT: If you experience a system that hangs (and doesn't recover after several minutes), please email support@cs.jhu.edu to let us know the name of the system that experienced the hang, the time of the hang, and what activity you were doing on the system at that time (which will help us troubleshoot the problem further.)
We will update here with our progress on circumventing this bug.
UPDATE: 1/29/20 We are applying alternative pre-bug kernels to those Linux clients that are available to be rebooted (to activate these kernels.) As more become available, we'll reboot them as well. We're hoping these alternative kernels will prevent the hangups until officially patched kernels become available.

PLANNED CS SERVICE OUTAGES

None at this time.


CS COMPUTING LAB AVAILABILITY

All CS computing labs available at this time.


AUDIO/VISUAL (A/V) OUTAGES

None at this time.