I guess my point was that even with all of the capabilities of out-of-band management, there's still some situations that require manual intervention. We only have 600 + servers, most of which have some type of out-of-band management (iDRAC, IP KVMs, etc.), but there are still those situations where a recurring hardware failure can render the server useless, regardless of our ability to power cycle the system.
- Sean On Wed, Oct 14, 2009 at 10:09 AM, Damien Solodow < [email protected]> wrote: > Considering those have a “virtual power button” that accomplishes the > same thing as pressing the real one, pretty unlikely unless there is a > physical power loss. The management cards also have their own network > interface which should be on a separate vlan from the servers themselves. > > > > But if worst comes to worst, you can have ip managed power units. So just > log into the PDU for the rack and reset the appropriately labeled power > receptacle. > > > > *From:* Sean Martin [mailto:[email protected]] > *Sent:* Wednesday, October 14, 2009 2:07 PM > *To:* NT System Admin Issues > *Subject:* Re: A look at the fully-packed racks inside a Facebook data > center facility. > > > > Hey chuck? Can you reset server “WEB23871? The iLO/DRAC card failed and > caused the server to bluescreen." > > > > Now what? :) > > > > - Sean > > On Wed, Oct 14, 2009 at 10:03 AM, Sean Rector <[email protected]> > wrote: > > +1 > > > > Sean Rector, MCSE > > > > *From:* Damien Solodow [mailto:[email protected]] > *Sent:* Wednesday, October 14, 2009 1:48 PM > > > *To:* NT System Admin Issues > *Subject:* RE: A look at the fully-packed racks inside a Facebook data > center facility. > > > > Walk to it? That’s what iLo/DRAC are for.. > > > > *From:* Jacob [mailto:[email protected]] > *Sent:* Wednesday, October 14, 2009 1:42 PM > *To:* NT System Admin Issues > *Subject:* RE: A look at the fully-packed racks inside a Facebook data > center facility. > > > > Hmm.. we have 80 servers in three racks in a room next to me. The web > servers are named “WEB01”, WEB02”, etc…. > > > > Hey chuck? Can you reset server “WEB23871”? > > > > Okay. Give me 30 minutes to walk to it… > > > > *From:* Sam Cayze [mailto:[email protected]] > *Sent:* Wednesday, October 14, 2009 10:21 AM > *To:* NT System Admin Issues > *Subject:* OT: A look at the fully-packed racks inside a Facebook data > center facility. > > > > Facebook Now Has 30,000 Servers. 25 Terabytes of Log Data – Daily > > > > > http://www.datacenterknowledge.com/archives/2009/10/13/facebook-now-has-30000-servers/ > > > > > > > > > > > > > > Information Technology Manager > Virginia Opera Association > > E-Mail: [email protected] > Phone: (757) 213-4548 (direct line) > {+} > > *Virginia Opera's 35th Anniversary Season <http://www.vaopera.org/>* *The > One You Love* > *Celebrate with a 2009-2010 Subscription: La > Bohème<http://www.vaopera.org/html/currentoperas/opera1.cfm>, > The Daughter of the > Regiment<http://www.vaopera.org/html/currentoperas/opera2.cfm>, > Don Giovanni <http://www.vaopera.org/html/currentoperas/opera3.cfm> and Porgy > and BessSM <http://www.vaopera.org/html/currentoperas/opera4.cfm>* > Visit us online at www.vaopera.org or call 1-866-OPERA-VA > > The vision of Virginia Opera is to enrich lives through the powerful > integration of music, voice and human drama > ------------------------------ > > This e-mail and any attached files are confidential and intended solely for > the intended recipient(s). Unless otherwise specified, persons unnamed as > recipients may not read, distribute, copy or alter this e-mail. Any views or > opinions expressed in this e-mail belong to the author and may not > necessarily represent those of Virginia Opera. Although precautions have > been taken to ensure no viruses are present, Virginia Opera cannot accept > responsibility for any loss or damage that may arise from the use of this > e-mail or attachments. > > {*} > > > > > > > > > > > > > > > > ~ Finally, powerful endpoint security that ISN'T a resource hog! ~ ~ <http://www.sunbeltsoftware.com/Business/VIPRE-Enterprise/> ~
