up_the_irons: anyone have any experience with the WD RE4 1TB drives? ***: Ehtyar has quit IRC (Quit: Going!)
LT has joined #arpnetworks mercutio: yeh
what abou tthem?
<tries to remember what had them>
they're like the wdc blacks with tler.
longer warranties
they're noisy
err longer mbtf
i think warranty time is the same
http://www.xbitlabs.com/articles/storage/display/1tb-hdd-roundup-4_6.html
their performance is very similar to the wdc black
with slightly higher max iops ***: hive-mind has joined #arpnetworks up_the_irons: mercutio: yeah mercutio: the only problem with them really is that they're a bit expensive
they were especially expensive just after the thailand flooding
like more so than seagate
that said for pure iops 15k sas disks are better
if you don't need lots of disk space
like the cheetah 15k.7 300gb disks aren't that bad
newegg (don't nkow many other places in the US) have a price roughly 2x for 1tb re4 versus seagate sas 15k
but like if you can take 8 hard-disks (say)
and make do with 1.2tb of data
you'll get better performance from the sas.
by quite a lot
and lots of cases don't take enough hard-disks to get iops up :(
you need a sas controller of course, but lots of people doing raid on lots of hard-disks use sas controllers anyway up_the_irons: yeah
I need the space though mercutio: 1.2 versus 1.6 million hours mtbf looks like up_the_irons: and big disk SAS, way expensive mercutio: oh you do
yeh
300gb is the sweet spot for sas
it really dependes if you need iops or space
the other thing you can do is you can short stroke
on sata disks up_the_irons: 1.6M MTBF is a joke if you ask me. I've never had a drive go longer than 30K hours mercutio: heh
howmany years is 30k hours? up_the_irons: a bit above 3 mercutio: oh
i've had disks go for longer than that
but what i've found is with some old disks once one dies others start dying up_the_irons: I've tried Seagate, Hitachi and WD. Maybe I'm doing something wrong ;) mercutio: hmm
well generally speaking people say you should have stable power and stable temperature
but how hot doesn't really matter unless it's extreme just consistent
but yeah disks fail
sas disks fail too :)
mm one gotcha now days is that lots of new disks do that 4k sector bullshit
which would be advantage of re4 up_the_irons: 4k sector bs? mercutio: yeh
they emulate 512 byte sector
by doing a read/write cycle
leading to really bad performance on some setups
you basically have to make sure all your offsets are 4k aligned
you hadn't heard of it? up_the_irons: no, i haven't
interesting mercutio: it's actually kind of annoying
if you replace raid set disks with 4k sector size disks you could very well run into performance degredation
the emulated ness makes it more likely ***: hive-min1 has joined #arpnetworks mercutio: some things support it better than others ***: hive-mind has quit IRC (Ping timeout: 245 seconds) mercutio: i think lvm has some stuff to do alignment etc
fdisk on linux now doesn't use the 1st megabyte of the disk when you do a new partition table up_the_irons: mercutio: lol, yeah i noticed that fdisk issue; through me for a loop at first... mercutio: up_the_irons: it took me ages to figure out how to fix it
you press c in fdisk
yeah i think it's 1 megabyte for ssds/flask etc
flash
which may have larger block sizes up_the_irons: mercutio: i try to use parted now
was annoying at first, but after you learn it, much easier to script stuff (which I need to do a lot) mercutio: well i normalyl just have partition straight on the underlying file system
but i had a "legacy" virtual machine
that had a partition table
i resized the partition...
then went to fdisk to delete and recreate
and it's like GAH
my problem with having a partition table is that you have to reboot to get the change
whereas if you just change the underlying volume you can get by without rebooting
and just do resize2fs
although don't know if that works in kvm
parted seemed to wnat to handle resizing itself.. i don't trust it :) up_the_irons: hah mercutio: well with things that could destroy data
i tend to like to take the familiar approach :) ***: bGeorge has quit IRC (Ping timeout: 272 seconds) mercutio: i suppose once you know it's safe you don't fret as much up_the_irons: yeah ***: bGeorge has joined #arpnetworks jlgaddis: up_the_irons: haven't really been looking since it first came back up, but haven't noticed anything. web sites are working and haven't gotten any alerts. up_the_irons: jlgaddis: roger that
My own VMs on kvr04 are also operating fine jlgaddis: i did a full offsite backup this (thursday) morning, i'm using ~4.2GB so it at least hit it disk pretty good for a few minutes while that ran up_the_irons: nice ***: hive-mind has joined #arpnetworks
hive-min1 has quit IRC (Ping timeout: 272 seconds)
hive-mind has quit IRC (Ping timeout: 240 seconds)
hive-mind has joined #arpnetworks toddf: up_the_irons: what I intended to say was that .. you used to stop the process entirely and start it anew, which caused a drop of connectivity, but now with HUP it reloads w/out a connection drop up_the_irons: toddf: cool.. i don't even remember doing that :)
but i suppose i did toddf: it is very clear in my mind because I don't have an expect like thing to signin again after it closes. so i had to manually notice and re-login. ***: hive-min1 has joined #arpnetworks
hive-mind has quit IRC (Ping timeout: 245 seconds) up_the_irons: mercutio: scroll down to "SmartAlign" : http://www.newegg.com/Product/Product.aspx?Item=N82E16822148840
mercutio: is that the 4k / 512 crap you were talking about? :) mercutio: yip
those disks are real fast for desktop use
but shit for raid
likve 170megabytes/sec fast
i've got 1tb in my desktop and 3tb on my file server
file server has openindiana which uses more space for small files on 4k sector size but does align properly
i have like 2tb free so i'm not too concerned
that said under windows it seeems terrible with multiple simultaneous copies
which although "generally" sucks.. some disks/filesystems deal better with than others
http://www.newegg.com/Product/Product.aspx?Item=N82E16822136533
i think these are still safe
i know the greens aren't safe
http://community.wdc.com/t5/Desktop-Portable-Drives/Is-WD2002FAEX-an-Advanced-Format-drive-or-not/td-p/186430
this seems to say which drives are 4k
but could be outdated, and it's only listing green
have you used wdc blacks with your raid controller before?
because some people have way more problems with the lack of TLER than others.
and you can't disable it on wdc blacks now days
err enable it up_the_irons: mercutio: ah good info
mercutio: supposedly the later 3ware 9650SE firmwares deal better with non-tler drives
and i always load the latest firmware
i may have a few blacks in my cluster
non-tler hasn't really been a problem that i can tell mercutio: well i'd go for a black then
they have 5 year warranty
i wouldn't go for a year long warranty disk
but for a desktop i'd recommend that first disk you linked.
the seagate
well for a desktop for someone who has a ssd :)
cos they're quieter than the blacks
faster sequential
(seagate has 1tb platter, i think wdc has 500gb)
well yeh if you'refine with non-tler, you're fine
you'd find out reasonably easily if you did have problems
(i've never had problems) ***: Webhostbudd_ has joined #arpnetworks up_the_irons: mercutio: the black's are not approved for raid though, afaik. so you can't rma them as easily.
i definitely wouldn't go for a 1 year warranty disk, that is just sad Webhostbudd_: no kidding
seagate is just sad mercutio: then go for re4?
it costs more though
we can rma blacks fien fwiw
forward replacement
which reminds me i have a black drive to rma :/ Webhostbudd_: mercutio: a black that went bad? up_the_irons: mercutio: so is the WD rma process decent? Webhostbudd_: up_the_irons: it was for me about 4 months ago
when i had a 640GB go bad
the only thing that sucks is that they discontinued that model, and ended up giving me a 750GB blue
I'm not really sure what they normally do when they have a current model which replaces it
it's pretty disappointing that I only have one hdd manufacturer to rely on, and some of their newer drives aren't so great..... up_the_irons: Webhostbudd_: sucks :( Webhostbudd_: i mean, who do we have left? wd and seagate
not trying to buy an enterprise drive from seagate with a 2 year warranty
that's rediculous
especially considering 90% of their drives get awful ratings and die all of the time
granted, ratings on drives are mostly RMA's so it is hard to really say which is more reliable up_the_irons: Webhostbudd_: we have tons of Hitachi's
i also got burned by Seagate Webhostbudd_: Seagate?
how did seagate burn them?
WD bought them...... up_the_irons: i mean we've had lots of seagates fail Webhostbudd_: oh
oh up_the_irons: our seagates have 5 year warranties though, i gotta get around to rma'ing some... Webhostbudd_: but the replacements will have 2 day warranties if i know seagate well enough up_the_irons: seagate probably learned after so many rma's, to drop their warranty down to 2 years ;) Webhostbudd_: yes they did
"we are looking to optimize our warranty policy" up_the_irons: haha ***: heavysixer has quit IRC (Quit: heavysixer) Webhostbudd_: they made some pretty funny statements when they made that announcement ***: heavysixer has joined #arpnetworks
ChanServ sets mode: +o heavysixer mercutio: up_the_irons: it's through your dealer normally
i got burned by maxtor
seagate and wd are basically the same
webhost: the blues are actually pretty good Webhostbudd_: maxtor,shiver
although
i can't complain all that much
i have some maxtors still running today and impretty sure they are 14 years old
i take that back, 16 ***: pcn has quit IRC (Ping timeout: 240 seconds) mercutio: heh
i had a really old scsi/esdi disk
which only worked if you left it on a while before using it
if you left it on it was fine
but if you turned it off and on again
(it was in external enclosure
it wouldn't work
but if you left it on for 20 minutes then turned it off/on it'd work
err if you turned it off and turned it on hours later that is
it was also bloody huge
and heavy
and only 300 megabytes
and only did like 700k/sec
but it stayed in that weird state for ages
and worked fine ***: fink has joined #arpnetworks
fink has quit IRC (Client Quit) mercutio: that disk must have been real ancient
ESDI was popular in the mid-to-late 1980s, when SCSI and ATA technologies were young and immature, and ST-506 was neither fast nor flexible enough. ESDI could handle data rates of 10, 15, or 20 Mbit/s
i don't think it went at 2mb/sec :)
hmm toshiba have new disks
i didn't know toshiba made desktop drives ***: heavysixer has quit IRC (Remote host closed the connection)
heavysixer has joined #arpnetworks
ChanServ sets mode: +o heavysixer
fink has joined #arpnetworks
ocm has joined #arpnetworks
ocm has left "Leaving"
ryk has quit IRC (Changing host)
ryk has joined #arpnetworks qbit: vps.. Y DNS NO WORK ON YOU ANYMORE?! ***: pcn has joined #arpnetworks Webhostbudd_: mercutio: toshiba makes drives!?
hahaha ***: LT has quit IRC (Quit: Leaving) Webhostbudd_: so apparently the toshibas are just hitatchis
soon to be wd's
it sucks that moving drive tech is pretty much dead, we need competition ***: gcw|mbpro has quit IRC (Remote host closed the connection)
jlgaddis has quit IRC (Ping timeout: 265 seconds)
HighJinx has quit IRC (Quit: Computer has gone to sleep.)
fink has quit IRC (Read error: Connection reset by peer)
fink has joined #arpnetworks
HighJinx has joined #arpnetworks
jlgaddis has joined #arpnetworks
jlgaddis has quit IRC (Changing host)
jlgaddis has joined #arpnetworks
gcw|mbpro has joined #arpnetworks
andol has quit IRC (Quit: leaving)
andol has joined #arpnetworks
Webhostbudd_ has quit IRC (Ping timeout: 272 seconds) up_the_irons: we need more nicks in here... invite your friends!
:) tooth: i could log in again. jdoe: I COULD SIGN UP FOR MORE VPSES TO IDLE FROM tooth: BUT WAIT YOU COULD INSTALL VIRTUALIZATION ON YOUR VPS
sup dawg. twobitha1ker: I could invite my bots ***: twobitha1ker is now known as twobithacker tooth: Heard you like virtualization. So we virtualized your virtual server, so you can idle while you idle. ***: heavysixer has quit IRC (Remote host closed the connection) up_the_irons: tooth: jdoe : LOL
this new server i'm configuring (AMD setup, and man it smokes) supports nested virtualization. I'll be offering beta accounts soon. tooth: so you can offer a virtualized server to someone, who can in turn, virtualize themselves... andol: Gimme! Gimme! Gimme! :) jdoe: sounds neat. ***: toddf has quit IRC (Ping timeout: 240 seconds)
heavysixer has joined #arpnetworks
ChanServ sets mode: +o heavysixer
toddf has joined #arpnetworks
ChanServ sets mode: +o toddf mnathani: up_the_irons, I just submitted an upgrade request for my VPS. No hurry though. milki: *wink* ***: Webhostbudd_ has joined #arpnetworks
Webhostbudd has quit IRC (Ping timeout: 250 seconds)
fink has quit IRC (Quit: fink)
heavysixer has quit IRC (Quit: BAMPF!)
gcw|mbpro has quit IRC (Ping timeout: 240 seconds)
easymac has quit IRC (Remote host closed the connection)
gcw|mbpro has joined #arpnetworks
dr_jkl has quit IRC (*.net *.split)
Webhostbudd_ has quit IRC (Quit: Leaving)
Webhostbudd has joined #arpnetworks
dr_jkl has joined #arpnetworks
gcw|mbpro has quit IRC (Ping timeout: 245 seconds)
fink has joined #arpnetworks
gcw|mbpro has joined #arpnetworks
HighJinx has quit IRC (Quit: Computer has gone to sleep.)
easymac has joined #arpnetworks
gcw|mbpro has quit IRC (Ping timeout: 244 seconds)
gcw|mbpro has joined #arpnetworks mnathani: up_the_irons: Any eta on my upgrade? ***: arenlor has joined #arpnetworks
fink has quit IRC (Quit: fink)
gcw|mbpro has quit IRC (Ping timeout: 252 seconds)
gcw|mini1 has joined #arpnetworks
HighJinx has joined #arpnetworks