IRC Logs for #crux Thursday, 2013-06-20

*** tjackson has quit IRC01:01
*** tjackson has joined #crux01:02
*** tilman has quit IRC02:04
*** tilman has joined #crux02:06
*** mavrick61 has quit IRC02:40
*** __mavrick61 has joined #crux02:41
*** myth has joined #crux03:23
*** tjackson has quit IRC04:20
*** deus_ex has joined #crux04:24
*** myth has quit IRC04:28
*** berban has joined #crux04:57
*** himynameisphil has quit IRC05:14
*** darfo has quit IRC05:27
*** darfo has joined #crux05:27
cruxbot[opt.git/2.8]: [notify] jre: update to 1.7.0_2506:59
*** ody has quit IRC07:03
*** sh4rm4 has quit IRC07:35
*** sh4rm4 has joined #crux07:35
*** acrux has quit IRC09:25
*** doomicide has joined #crux09:25
*** acrux has joined #crux09:38
*** Guest33475 has joined #crux09:52
*** ody has joined #crux10:24
*** nullmark has joined #crux10:38
Romsterclang has 404
Romster is that it? days Clang source code (9.4M)11:34
Romsteryeah it is confirmed from the arch pkgbuild11:37
Romsterwhy would they rename it11:38
jaegeremail them and ask12:12
Romsterprobably wouldn't listen to a nobody like me12:13
Romsterthey probably plan to rename it or integrate it somehow at a guess.12:14
Romsteron another note i'm finally recovering12:14
frinnstwat, you were broken?12:15
frinnstmanflu is the worst12:15
Romsterlets just say over 20 visits to the toilet, aching joints, headache. hazy thinking12:15
Romsterstarved myself too12:16
jaegerDo what I did, get swine flu. After that you're very flu-resistant. :)12:16
Romsternow i can actaully keep food down and not that far down the other end12:16
Romsteri went and ate 600mills of soup with 2 and a half teaspoons of cayenne hot pepper, that fixed my gut up.12:17
Romsteri would not recomend doing that though -_-12:17
frinnstyeah ought to kill whatever is in there :)12:18
Romsterif it don't kill yourself in the process12:19
Romsteruneven heart rhythm from it12:19
frinnstyeah that shit is dangerous. a friend of mine died from heart failure ~6 months ago12:20
frinnstpreviously healthy, just dropped dead pretty much12:21
Romsteryeah, that's why i don't recomend it12:22
Romsteri've had it before when i had far too much caffeine too12:22
joacimi hear there are people that have died from those cinnamon challenges12:22
joacimtheir lungs collapsed12:22
frinnstyeah but thats just darwin in action12:23
Romsterwhat on earth were they doing?12:23
Romsterinhailing it?12:23
frinnstbeing stupid for facebook12:23
frinnstyoutube it12:23
joacimeat a spoon of the stuff12:23
Romsterfuck that shit, that's asking for trouble12:23
joacimi just put it on my porridge12:24
Romstermakes more sense too do that12:24
Romsterlove cinnamon12:25
Romsterwhat's next sniffing up nutmeg?12:25
joacimfirst time i've heard of nutmeg12:26
Romsterreally it's used in buns12:27
frinnstthere's a well known "underground" forum in sweden called flashback. a guy posted there after he ate a bunch of nutmeg. he was out for about 24hrs12:27
*** pitillo has quit IRC12:28
RomsterNutmeg was extremely popular in much of the world from the fifteenth to the nineteenth centuries (Herbst 2001) and remains widely used today as a spice for food and drinks. As a culinary spice, with its unique aroma and taste, nutmeg has added to the human enjoyment of many foods. The essential oil and dried nut also are employed medicinally; however, the effectiveness of treatment versus the toxicity of overdoses makes its medicinal use generally no12:29
Romstert recommended. The risks associated with consuming too much nutmeg reflects on the importance of being informed regarding even easily obtained food items.12:29
Romstersilly guy12:29
frinnstyeah.. darwin awards12:29
*** pitillo has joined #crux12:30
joacimturns out i do know what nutmeg is, we just call it something else here.12:32
nogagplzwhat do you call it there12:34
Romsternice spice but never over do it12:37
frinnstmuskot, almost jaeger :)12:38
jaegerIt was a google-based guess :)12:39
*** Roomster has joined #crux12:39
*** Roomster has quit IRC12:39
*** Roomster has joined #crux12:39
frinnstlooks like it might be called that infinish12:40
frinnstin finnish asdf12:40
jaegerI was thinking for some reason joacim is in finland12:40
frinnstno norway12:40
jaegerah, ok :) I'll go sit in the corner12:40
joacimmy cousin buys his beer in finland, so it is close enough i guess12:40
frinnstthis is finland <- Not really safe for work / lunch12:41
jaegerI will avoid it for now as I'm at work and have not had lunch yet12:41
jaegerUnrelated to spices, I've been seeing a really weird issue with an ubuntu box at work12:45
jaegerit's running 12.04 so the desktop is unity/gnome312:45
jaegerit has an nvidia video card running the proprietary drivers12:45
jaegerThe problem is this:12:45
jaegerFor periods of time that might last 10 seconds or 5 minutes, starting a new terminal (and ONLY terminal) will freeze the desktop display for about 5 seconds12:45
joacim <- more finland12:46
frinnstnice, cpu spike?12:46
jaegerI wonder if it's the desktop compositing or vte or something else entirely12:47
jaegernot sure, I'll have to watch it next time it's happening12:47
frinnstjoacim: brilliant12:47
jaegerI've got no idea what triggers it to start or stop, so bizarre12:47
jaegerIt doesn't happen on the very same machine if booted into another distro, something about ubuntu's setup or unity itself12:48
frinnstweve started to see some strange stuff with vmware the last couple of days12:48
frinnstsnapshots freezes the vm completely for a few seconds.. pings timeout and crap12:48
frinnstand only on linux vm's12:49
jaegerthat would suck, definitely. new behavior?12:50
jaegerDo you mean vmware snapshots or SAN snapshots?12:50
frinnstnothing has really changed either12:50
frinnstduring veeam backups12:50
frinnstit seems to be when the snapshot is deleted12:50
jaegerinteresting... does veeam do deduplication of the snapshots?12:51
jaegerI've no experience with veeam myself12:51
frinnstthere is a veeam proxy vm that does all the deduplication and processing. so, veeam creates a snapshot, the proxy reconfigures itself and attached the targets disks and process those12:52
jaegerIs there any chance the proxy vm creates some kind of network bottleneck?12:53
jaegerOr I/O even12:53
teK_do you have a ratio wrt space-saving effects of the deduplication? I imagine it to be useful for OS-data almost exclusively12:53
frinnstnah, the issue is only when the snapshot is released i think12:56
frinnsti started to look at it before but got distracted with stupid users12:56
jaegerWell, the reason I asked about dedup is because snapshot deletion ended up being what killed all my VMs for 2 months back when we were deduping our ZFS SAN12:56
frinnstyep :)12:57
jaegera long shot but that's the first thing that came to mind, hehe12:57
frinnstthis must be either some new vmware bug or perhaps some bug with open-vm-tools12:57
frinnstwe run that on most of our linux guests12:58
jaegerI've seen weirdness with open-vm-tools now and then with startup/shutdown/quiescing the filesystem, but never freezes12:58
jaegerI tend to use the official tools everywhere I can now12:58
frinnstyeah but that's painful with kernel updates12:58
jaegeractually the last holdout on that was the github enterprise VM and they've changed their minds on that recently12:58
jaegerI don't mind, myself, but you might have a bigger deployment than I do12:59
jaegermost of the time selecting the hosts in vsphere and doing an automatic tools upgrade works12:59
jaegerer... selecting the VMs, not the hosts12:59
frinnstbefore booting the new kernel?13:00
jaegerno, after13:00
frinnsthm, last time i booted a new kernel vmware-tools would not start13:00
frinnstrhel ~6months ago maybe13:00
jaegerI've only got one rhel VM so not sure there... most of mine are ubuntu server and handle it well13:01
frinnstmight want to revisit that then13:01
frinnstim unclear about vmwares position on their tools for linux.. all appliances from vmware use open-vm-tools13:02
frinnstthey usually run on opensuse13:02
jaegerinteresting, didn't know that13:03
frinnst--- ping statistics ---13:04
frinnst16 packets transmitted, 0 received, 100% packet loss, time 14999ms13:04
frinnstits removing the snapshot right now..13:04
frinnstteK_: not sure exactly what is dedup and whats compression, but the backup set that contains a bunch of 2008r2 vms are 2,6TB in total and a full backup uses ~830GB13:07
jaegerlooks like the CPU does spike on a single core when that terminal issue is happening13:08
jaegerI think it's compiz but hard to tell as the terminals with top running freeze too :) I'll try watching htop from an SSH session :)13:08
frinnst64 bytes from icmp_seq=134 ttl=64 time=1.08 ms13:09
frinnst64 bytes from icmp_seq=176 ttl=64 time=1.57 ms13:09
frinnstdropped quite a few pings again..13:09
teK_either way .. nice ratio13:09
frinnstfucking depressing13:20
bloouupKind of off topic but does anyone know of any actively maintained linux distributions that use plan9port instead of the gnu coreutils?13:21
jaegerNot I13:23
bloouupbecause that would be so truly lightweight13:25
cruxbot[compat-32.git/3.0]: glib-32: 2.36.2 -> 2.36.313:42
cruxbot[compat-32.git/3.0]: gtk-32: 2.24.18 -> 2.24.1913:42
cruxbot[compat-32.git/3.0]: llvm-32: 3.2 -> 3.313:42
*** dkoby has joined #crux13:55
jaegernogagplz: bluez in the emulators repo needs a .md5sum update and seems to be missing a libical dep14:08
frinnsthm, seems we found a workaround for the freezing problems during backup14:19
frinnstchanged the transport mode from the hypervisor io stack to the network stack14:19
frinnstslower transfers but no more total lockups14:19
jaegersounds prefereable14:20
jaegerpreferable, too14:20
frinnstit is quite a bit slower14:20
frinnstfrom up to ~150mb/s to ~30mb/s14:21
frinnstoh well, as long as no servers are affected..14:21
jaegerslow is probably better than unreachable :/14:22
frinnstyep :)14:23
frinnstbut still dont get why we started seeing this issue in the last couple of days14:24
*** Kaishi has quit IRC14:24
frinnstits been running just fine for a year or so and for about a month in our new environment14:24
jaegeryeah, that's odd... did anything add a lot more I/O in the recent days?14:24
frinnstthe io is pretty calm14:24
teK_- Move VM to a host with more available resources.14:24
teK_I always like that suggestion14:24
teK_performance sucks? Well.. throw $$$ at it!!1114:24
frinnst120gig ram available and cpu is pretty calm too14:25
teK_while in the first place $$$ were thrown at better using the _available_ performance14:25
teK_more manufacturer madness: audit people of our ERP system want to know (for some users) what they are doing with the system14:26
teK_they use the fucking license. The rest of the story is none of your business.14:27
frinnstlooks kinda sexy14:28
teK_2. SMB has better performance than iSCSI? Would not have expected that.14:29
frinnstiscsi has locking issues iirc14:30
frinnstsmb < iscsi < nfs14:31
teK_I could not belive NFS was used for storage + vmware first place14:32
frinnstwhy? its awesome :)14:32
frinnstwe used to run iscsi.. nightmare14:32
teK_it has shitty behaviour if the serve gets lost14:32
jaegerI've used a little of both with our vmware stuff, much rather use NFS14:33
teK_great; will remember :)14:33
jaegerthe VMs can shit themselves if the NFS server goes away (like it did with our ZFS dedup issues)14:33
teK_v3 or 4?14:33
frinnstvmware uses 314:33
jaegeriscsi had some annoying limitations, though, like 2TB max14:33
jaegerThough as far as I know they've fixed that with a VMFS upgrade14:34
frinnstthere are workarounds iirc, but the 2tb limit is still there14:34
jaegernow that we know never to use dedup in production, NFS has been rock solid14:34
frinnstwe use nfs because it gives us *better* dedup :)14:34
jaegerwell, I'm talking specifically about ZFS dedup14:35
teK_so what's the difference in your setups? I guess both of you use ESXi 5 and vCenter?14:35
jaegerfor reference, 128GB RAM is *NOT* enough to handle the dedup tables for 37TB of data14:35
frinnstwe use a lot less data, maybe ~20tb in total14:35
jaegerthe data we were deduping isn't even vmware-related but ZFS dedup tables are pool-wide so performance was still trashed for vmware14:36
frinnstwe are migrating away from our old san (fas 2040) to our new (fas 2240), with new blades, 10gbit switches etc14:36
frinnstand from our physical cisco firewalls to vmware vshield14:37
jaegerWe're split between two SANs currently... oracle zfs storage is one and the other is an EMC Isilon X200 cluster14:37
jaegerall the vmware stuff is on the isilon cluster, love it14:37
frinnstthe old san uses iscsi and the new filer uses nfs14:38
teK_are there systems you won't move into the vmware infrastructure due to (data) safety concerns?14:40
jaegerNot in my case, no14:41
*** vaddi has joined #crux14:46
jaegerAt this point the only critical servers we have not virtualized are the exchange hubs and our RSA appliance14:47
jaegerthe RSA appliance we could get a virtual version of but we haven't bothered14:48
jaegeras for exchange, nobody knows what the windows admin is thinking14:48
jaegerI don't think they're even backed up14:48
teK_s/what he's/if he's/g ;D14:51
jaegerI can't/won't argue that point, hehe14:51
*** ody has quit IRC15:07
*** dkoby has quit IRC15:14
*** toumbia has joined #crux15:21
*** crazyhors has joined #crux15:49
*** heroux has quit IRC16:05
*** heroux has joined #crux16:05
*** Roomster has quit IRC17:09
*** pitillo has quit IRC17:10
*** pitillo has joined #crux17:11
*** jdolan_ has joined #crux17:37
*** jdolan has quit IRC17:37
*** vaddi has quit IRC19:25
*** doomicide has quit IRC19:36
*** doomicide has joined #crux19:38
*** ody has joined #crux20:55
*** lasso has joined #crux21:28
*** Rotwang has joined #crux21:36
*** Rotwang has quit IRC21:45
*** lasso has quit IRC21:50
*** doomicide has quit IRC22:27
*** pitillo has quit IRC22:28
*** pitillo has joined #crux22:34
*** y3llow has joined #crux22:41
*** irclogger_ has joined #crux23:27
*** Goodbox has quit IRC23:55

Generated by 2.11.0 by Marius Gedminas - find it at!