#

Monthly Maintenance: December 1st 2014

Our next regularly scheduled planned maintenance event will take place on December 1st from 7AM-11AM.

Tasks:

**We will be setting a default time for all SLURM partitions (including the interactive partition) of 10 minutes.

Use the -t flag to set a time. If you do not include a specification of how much time your run requires, it will now by default get 10 minutes. If you do not know how much time your code will take, either schedule a test run to find out, set it to your best guess, or set it to be the maximum partition time.

We are making this change to improve the backfill scheduler; accurate estimates of the run time make the scheduler more efficient. Note: this is not a change to the maximum time permitted for a partition, only the default setting for a run that does not specify any time requirements.

For further information on batch commands and how to implement -t please see here.

**As planned and publicly announced, we will be enforcing file retention policy on both the regal and holyscratch filesystems.

**As part of our regular efforts to maintain appropriate and up-to-date versions of software and operating systems, we will be upgrading our login nodes (rclogin01-rclogin12) from Centos version 6.4 to 6.5. There should be no noticeable difference to you.

**In order to improve our networking services, we are replacing one of our firewalls with a newer unit. Since a number of customer virtual machines reside on a network which has to move to this new firewall, we will need to temporarily shut down and restart the following list of virtual machines. We are reserving a two hour window to keep these systems offline, but we expect the network move to take much less time than that.

List of virtual machines that will need to be temporarily unavailable on December 1:

account
account-dev
askbot
astrobites
bicep
bicepkeck
billy
biobakery
bussle
catmaid
ccfa01
ccfa02
ccfa03
celeste
cgrweb-prd
chopchop
cnl
cohenweb
collections
daschprivate
docker
dokken
downloads
eggan-int
eggan-wp
emattison1
eqtl2
evolmeta
fs-survey
git
giza-web
gmarnellos1
haep
hank
helium
hidelab
hulssc
illustris
informatics1
informatics4
informatics5
in-pn-interactions
ip2-ext
itssandbox
jbaker
kerbslap
kuangwiki
lieberwp
loni
maelstrom
mczweb
mczweb-dev
molspaceweb
msprl
murray
murraylabweb
neuroinfo-dev
nocera
nrgicog1
nrgicog2
obs1
octest
odybot
oshealab
oshea-wiki
owncloud
particlebites
phylopsy
physmt
portal
ptolemaios
qiime
ramanathan-vm
rcwebsite2
reuven1
rinnlab
ruiqingcao
scc-prod
sccr-dev
scc-stage
shiny
socialmol
sptclusterarchive
status
stembites
structure
survana-store01
synccp
telescopes
tpp1
variantviz
vcg
visalam
waywiser
webgeneric
xrayq
z-brain

We are doing no work on either compute nodes or storage. SLURM will continue to process jobs. Access may be limited due to the login nodes upgrade, but it will be a rolling upgrade - the login systems will not all be down at the same time.

Please contact us with questions/requests. As always, thank you for your cooperation.

CC BY-NC 4.0 This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. Permissions beyond the scope of this license may be available at Attribution.