Question for the group, given the significance of this service, would you like to move this to DivIT systems which are patched and maintained? If so, I can investigate.
Very Best Regards, Emery Rudolph, MS Manager IT-ETI-PS Enterprise UNIX Services University of Maryland (301) 405-9379 http://www.umd.edu -----Original Message----- From: UM Linux User's Group [mailto:UM-LINUX@LISTSERV.UMD.EDU] On Behalf Of Ken Tossell Sent: Thursday, July 31, 2014 4:33 PM To: um-linux Subject: Re: [UM-LINUX] Mirror.umd down? The server had two 36 GB WD Raptors (10k, 3.5", SATA II), one of which stopped working, for the OS, some static files and the mirroring scripts. I think these disks were less than half full. There are four 2TB Seagate Barracuda XTs connected to the faulty SATA controller (a 3ware 9650SE). We have about 1.8TB free (of 5.4TB). We have two spare disks, but we're out of drive bays. I believe the SATA/RAID card needs PCI Express x4. Ideally we'd replace that card with another SATA controller and either a) replace the Raptors with SSDs or b) replace the Raptors with working magnetic drives while adding a dedicated SSD for caching purposes. Ken On 07/31/2014 04:16 PM, Yehuda Katz wrote: > What is the current HD usage? > I am a current student (PhD, so I will be here for a while) and I have > some spare hardware if necessary that we could use to revive it. > Having a local mirror is really helpful. > > - Y > > > On Thu, Jul 31, 2014 at 3:26 PM, Ken Tossell <k...@tossell.net> wrote: > >> The mirror server's had a combination of a failed OS disk, some >> filesystem trouble and a broken SATA controller. We're not sure why >> it locked up last Friday, but it's down right now because the SATA >> controller had its very frequent intermittent failure after the >> server was brought back up yesterday. We hoped the controller would >> start working after a remote reboot, but the machine never came back online. >> >> Even after we get the server running, there's a good chance of >> downtime with the current hardware. We are running with zero >> redundancy in the / disk, and the remaining drive is identical to the >> one that failed. Also, the SATA controller could break and disconnect >> the main software storage disk array at any time. There's a ~90% >> chance that the array won't come up after the next power outage. If >> that happens, we'll need to manually disconnect and reconnect the >> power several times in order to get the server going. >> >> I do some maintenance (adding mirrors, mostly), but since I'm no >> longer at the university, other people have been taking care of all >> of the issues that need local attention. >> >> Ken >> >> On 07/31/2014 01:33 PM, Yehuda Katz wrote: >>> I believe the mirror is run by Ken Tossell (k...@tossell.net). >>> >>> - Y >>> >>> >>> On Thu, Jul 31, 2014 at 12:13 PM, Moshe M. Katz <mmk...@umd.edu> wrote: >>> >>>> Hello all, >>>> >>>> There's been very little traffic on this list recently (correction, >>>> none at all), so I hope that people are still around. >>>> >>>> The mirror.umd.edu server appears to be down, and the only >>>> published email address is "mirror@hostname", which is, of course, down >>>> with it. >>>> >>>> Does anyone know who the current maintainer is and/or what the >>>> story is here? >>>> >>>> Thanks, >>>> Moshe >>>> >>>> -- >>>> Moshe Katz >>>> mmk...@umd.edu >>>> (301) 867-3732 >>>> >>>> >>>> >>>> On Thu, Jul 31, 2014 at 11:56 AM, Jeff Stuckman <stuck...@umd.edu> >> wrote: >>>> >>>>> Do you know who runs this mirror? >>>>> >>>>> >>>>> >>>>> *From:* Nicholas Rossomando [mailto:nross...@terpmail.umd.edu] >>>>> *Sent:* Thursday, July 31, 2014 11:55 AM >>>>> *To:* Jeff Stuckman >>>>> *Subject:* Re: Red Hat: What we've found so far >>>>> >>>>> >>>>> >>>>> Well, right now, zero data from UMD. The mirror is down and has >>>>> been >> for >>>>> several days now. If it comes back up, it's definitely ideal for >> however >>>>> far back it goes. >>>>> >>>>> Noted, and will do. >>>>> >>>>> >>>>> >>>> >>>> >>> >> >> >
