builder drops 6 cat5e pairs in the server room from the call center floor. all of them blue (we normally use blue for data and white for phone). 2 of them arent long enough to reach the patch panel rack and i'm all out of 110 blocks for the phone side. wtf friday.
@ColdT Well, the clinicians tend not to know which side of the keyboard to bang on, and they make the hiring decisions for their peers. So I'm not surprised.
facepalm Something in Broadstripe's network mgmt suite is resetting our staff cable internet connection down to 4Mbit/s every night. I've spent 2 hours on the phone each morning this week getting it reset back to 15.
@JeffFerland Agreed. Well, there's nothing at all wrong with the brand new fax machine. I'm pretty sure their software between their ears is defective though.
@JeffFerland Because it's the most advanced technology that mental health professionals can figure out? I still get requests from nurses, doctors, and managers to have a personal fax machine in each office.
@voretaq7 But you see, that would involve using a computer. They don't WANT to use the computer. The fax machines allow them to avoid that particular unpleasant reality.
In telecommunication, a ringer equivalence number (REN) is a somewhat arbitrary number which denotes the electrical load a telephone ringer has on the line. In the U.S., this is determined in accordance with ANSI/TIA-968-B (August 2009).
REN is a United States-developed yardstick, but analogous systems exist internationally. In some countries, (particularly in Commonwealth nations), the REN is better known as the ringer approximated loading number (RAL).
Explanation
A ringer equivalency number of 1 represents the loading effect of a single "traditional" telephone ringing circuit, such...
I haven't even thought about that term in five years.
@BartSilverstrim I'm a NorthWesterner that screwed up and lived east of the rockies for five years before escaping (barely with my life) and now live in Phoenix AZ with all the citrus trees, hummingbirds, peach faced lovebirds, horse farms just down the road and oh yes NO SNOW.
@Adrian Yeah, he's talking to HR right now. You're fired.
@BartSilverstrim Well, there's republicans, guns, gang-related home invasion mass murders, and Starbucks on every corner. Does that help?
I have never heard of "Areca" brand RAID hardware. Until a client's server refused to boot up last night.
I guess I should thank them for billable time, but I am not impressed with what I've seen of them so far.
"Areca designs their own custom ASIC, a high-speed IDE controller, and RAID 6 engine on their controller to achieve high data throughput and availability."
Interesting. Not many RAID controller manufacturers can say that these days.
"Areca. Customized failure for all your downtime needs."
I used to drink straight rum, then I took an arrow to the knee after stumbling onto an archery field and threatening the competitors with projectile vomit.
cook the bacon - cook some hash browns with peppers and onions in it. wait till both are done - beat 3 eggs into some milk - pour into frying pan - scramble like normal but while liquidy, put 1/4 of the hash browns into it and crumble 3 strips of bacon into it. put between 2 slices of buttered toasted white bread.
@Adrian I converted KVM VMs using VMWare converter.
@Iain No, THEY drew on whiteboard :)
Some interesting questions that I hadn't had before.
"imagine you have a server... Let's say through a company acquisition... What would you do in order to determine what it does? Walk us through the steps..."
@ewwhite First, Cheeze Whiz. Second, a canister of helium. Third... do you have any cats?
Seriously though: Windows: appwiz.cpl, sc.exe, nmap that sucker from the outside *nix: dump package manager contents, ps aux, nmap that sucker from the outside
But that was an example. Other questions were interesting...let's say you have an app that's dumping data to do disk, but pauses frequently... What do you do?
@ewwhite I'd start with iotop and see which processes were writing, and then pausing. From there, it's totally up to which processes are involved. Plus, if the filesystem is ZFS, it's supposed to do that. There's a "breathing" action to writes that's normal.
Windows: Use performance graphing to track the process. PRetty simple in 2008+. Then, based on the process, you can move forward.
I'd say CentOS or Scientific Linux with a realtime kernel compiled from Red Hat MRG source RPMs or using the CERN pre-built binaries.
See: Compiling realtime kernel from RHEL 6 MRG sources on CentOS 6
and CentOS 5 - realtime patch
The caveat is that simply installing a realtime distro won't a...
@ewwhite You got one downvote, got upvotes from me and someone else. I've gotta finish reading some of the stuff you linked to, even though I don't need real time, I like it.
Bacula won't make use of 2 tape devices simultaneously. (Search for #-#-# for the TL;DR)
A little background, perhaps.
In the process of trying to get a decent working backup solution (backing up >20TB ain't cheap, or easy) at $dayjob, we bought a bunch of things to make it work.
Firstly, ther...
I don't have an autochanger (we do everything to disk, and i want my backups to be sequential to files), but I'd be interested in knowing if this is a known bacula deficiency...
2012-06-08T19:16:08.235Z cpu1:2049)WARNING: ScsiDeviceIO: 1218: Device naa.600601601fd02d0046b558e32553e111 performance has deteriorated. I/O latency increased from average value of 4622 microseconds to 106873 microseconds.
and we use AppAssure Replay - backup's stand alone servers to a VM on ESXi so incase of catastrophic hardware failure, the VM can be turned on and take its place. that, is actually pretty cool and we like it. But they were recently bought out by Dell :(
One more interview question... Imagine you work at a online music startup and your predecessor cataloged music using the title of the audio files... So you end up with a poorly scaling mess of files with typos.
@ewwhite Files, this crap is straight on the filesystem? The metadata for music files is on the effing filesystem? Okay, first you go all @Chopper3 and bleed out the CTO's family in front of him for allowing this.
Step 1: Hard-Link all the files to their primary key name (presumably an integer) Step 2: Tell the developers to fucking man up and use the primary key integer id as the filesystem name or you'll shove their testicles into their eye sockets. Step 3: Shove developer's testicles into their eye sockets. Because they deserve it. Step 2:
yeah, but name as a file name isn't nessecarily bad. we've had 2 digital voice recorders in this office. the first one named the file name with date, employee name, time, channel on pbx, etc... but the server sucked. was hard to find anything. but, if the server failed, it was easy to visually identify a file. the new one had an index for the file name and the meta was stored in mysql. one day the DB failed and you couldn't tell 1 file from another. but the software was better.
I have a number of CentOS 5 servers which use 3ware RAID controllers.
These servers are bugging my team with messages about minor temperature changes, like this:
Jun 8 12:32:39 HOST smartd[1231]: Device: /dev/twa0 [3ware_disk_01], SMART Usage Attribute: 194 Temperature_Celsius changed from 119...
Okay, I have a client who's dedicated server at 1and1 suffered a HD failure. RAID 1 was in place, so I'm making sure things are backed up before having a second drive placed in and rebuilding the array.
Then I saw this little bit of loveliness on the remaining good drive:
smartctl -A /dev/sda
Current Drive Temperature: 30 C Drive Trip Temperature: 25 C Manufactured in week 30 of year 2002
@WesleyDavid But the other issue was that it seems like there were 10 departments handling IT... Team A installs the OS. Team B copies the OS. Team C owns the servers... Team D manages production.
<troll your helpdesk staff> go to English dept. staffroom... transpose the M and N keys on all the staff's keyboards because... fuck them. </troll your helpdesk staff>
Just a quirk of mine based on completely subjective experiences concerning drive failure. (That research paper you linked to would probably give me more cogent thought on the matter though!) I don't like drive failures as unplanned rebuilds have disappointed me in many ways. So I like to cycle drives if it's practical.
For lots of DAS, it's not practical and keeping track of hardware like that is nigh of impossible. NAS / SAN storage is easier to live that out with.
I like to have array rebuilds as part of planned maintenance rather than on an as needed basis, because "as needed" means "when bad stuff happens" and that's just uncomfortable no matter how well your array handles failed drives.
Okay Seattle, stop calling me and not responding to my hellos.
@WesleyDavid I prefer good backups -- unplanned rebuilds have given me unpleasant surprises too, but it seems that my chance of a failure during a rebuild is roughly equal at any point after the 2-3 year mark, and the marginal costs of cycling disks are too high compared to a failure of a redundant system for 24 hours...
@ewwhite @Iain I've been getting calls from the NW a lot in the last two months. Portland, Seattle, Vancouver WA. My number must be close to some business number I guess. I answer in my typical business fashion and ask "how can I help you?" and they immediately hang up.
@voretaq7 Yes, good backups are boffo. I just hate, hate, hate needing to use them for anything other than tests. The arrays I deal with aren't terribly expensive, so that probably factors in too. If I was on NetApp or EMC stuff, I'd leave well enough alone since 1) Warranty, 2) That stuff is solid RAID 01111561 anyway.
@Zypher I can sway the judge. "Your honor, if you accept my argument that users are not people and therefore shooting them is not murder you can extend it logically to frivolous litigants who don't understand that the judiciary is not just a big room for them to stand in and whine..."
Nice. The QEMU-IMG converter in Centos6 worked like a charm. Converted my raw disk images to VMDK and VMWare fired them right up when I hooked a new VM up to them.
It's kinda strange when stuff in Linux just works without any fiddling first.
A drive is beginning to fail and I only know the device by its /dev/sdb device file designation. What are the ways that I can use to correlate that device file to an actual hardware device to know which drive to physically replace?
Bonus: What if I don't have /dev/disk/ and its sub directories o...
But if you were just talking about raw SCSI disks, then you're talking about a software RAID scenario? If so, then the smartctl, /proc/scsi/scsi and hmm...
Not sure. My last-bad boss did software RAID everywhere and I ended up having to look for a failed disk... I'm trying to recall what I did.
I think Ethabelle went to look at the server for lack of LED activity.