[05:46] <FarhaadN> hi every one, i use ubuntu server 12.04 , i need upgrade openssl from version 1.0.1 to version 1.0.2 , any issue happend? what dependency need too this
[05:46] <FarhaadN> only way to upgrade is from source?
[05:46] <FarhaadN> apt-get dont upgrade to version 1.0.2?
[05:52] <FarhaadN> no one answer?
[06:01] <FarhaadN> pleaseeee
[06:01] <FarhaadN> i neeed this
[06:07] <FarhaadN> hi every one, i use ubuntu server 12.04 , i need upgrade openssl from version 1.0.1 to version 1.0.2 , any issue happend? what dependency need too this
[06:07] <FarhaadN> only way to upgrade is from source?
[06:07] <FarhaadN> apt-get dont upgrade to version 1.0.2?
[06:25] <nasko_> hi
[07:33] <lordievader> Good morning
[10:48] <aderyugin> hi, folks! I have a question regarding packages for OpenStack Newton. Is there any way to get murano packaged for newton?
[10:48] <aderyugin> jamespage: ^^
[10:49] <jamespage> aderyugin, possible - we normally sync that via debian
[10:50] <aderyugin> jamespage: ok, is there any estimates for package sync?
[10:50] <jamespage> aderyugin, just synced it - will go into staging today
[10:51] <aderyugin> jamespage: thanks
[10:55] <ram___> Hi. Can I install OpenStack Liberty using the OpenStack Autopilot 15.01?
[11:24] <warpx> Hello
[11:24] <warpx> Does anyone over here have Thunderbolt networking experience?
[11:24] <warpx> I'm looking into using Thunderbolt in the same rack to replace NIC
[11:26] <patdk-lap> not sure how you can do that
[11:26] <patdk-lap> thunderbolt is not a network, it's pcie
[11:27] <warpx> Yes, I know
[11:27] <warpx> However, it's PCIe, so I should be able to do everything and anything with it
[11:28] <patdk-lap> heh?
[11:28] <warpx> All I need is to get TCP/IP over it
[11:28] <patdk-lap> pcie is point to point
[11:28] <warpx> Yes
[11:28] <warpx> I failed to find switches
[11:28] <patdk-lap> they make pcie switches
[11:29] <warpx> Yes
[11:29] <patdk-lap> the issue is, only one side can control the switch, so not useful
[11:29] <warpx> Yes
[11:29] <warpx> I was hoping to daisy chain my db hosts
[11:29] <patdk-lap> you need to start making your own asics
[11:29] <patdk-lap> and drivers
[11:29] <warpx> Yes
[11:29] <warpx> I was hoping someone started this
[12:00] <trippeh_> intel has some "multi host" pcie network stuff. but its gonna cost ya
[12:01] <warpx> trippeh_: I'll look into that
[12:02] <trippeh_> I think the general idea is to use it in something like a blade chassis.
[12:02] <warpx> So I need rackspace
[12:02] <warpx> For blades
[12:02] <warpx> hmm
[12:04] <Mr_Pan> i tried upgrade 15.04 >> 16.04 ... but now i have the  message "The user `syslog' is already a member of `adm'" . The installation is blocked. Any ideas?
[13:25] <dr4c4n> hi, can anyone give me a hand with bridges and vlans? I have created a bridge, and the vms that I have created in kvm can talk to one another, but if I try to add another physical box on the same vlan, they can't communicate.
[13:25] <dr4c4n> they can't communicate to the physical box <clarification
[15:31] <darsparx_> anyone here know how to trouble shoot msmtp? It's ignoring my config file so I can't even finish setup to use gmail with my DO server...
[15:33] <rattking> did you try -v to "Print lots of debugging information"
[15:34] <darsparx_> no but i am noticing the fact it' ssaying permission denied now which is weird....wonder what usergroup it's trying to access it at now that i think about it
[15:35] <rattking> there is also -P to print the configuration settings that would be used
[15:41] <darsparx_> well I think I got it to work sorta fixed it....though looks like I have to wait for my mx record to fully go through >_<
[17:02] <darsparx_> well now i've gotten msmtp to work on the server(16.04) but can't seem to get php to use it, it just keeps going for the error side of my test, and there's nothing in my specified log file... :-\
[20:08] <tomreyn> hi
[20:09] <tomreyn> i'd like to move vm's over the internet between two DCs. those are qemu (KVM) linux guests, stored on (LVM2) LVs.
[20:09] <tomreyn> so far the only solutions i found for copying the entire partition from the source HV to the target HV is a combination of dd and ssh. this seems both quite ineffecient and, moreover, error prone.
[20:10] <tomreyn> i would expect that if the network link fails / times out for some reason then i 'd need to restart the copy from the beginning?
[20:10] <nacc> tomreyn: are you using libvirt? i have no idea if it's doable for your configuration (or advisable), but you could do an offline migration
[20:10] <tomreyn> is there another approach which provides resumable transfers, and better compression than ssh -c?
[20:11] <patdk-lap2> what is the connection speed between the two?
[20:11] <patdk-lap2> not likely to have resumable at all
[20:11] <patdk-lap2> I would just dd | pigz | mbuffer
[20:11] <tomreyn> yes i'm using libvirt. i tried an online migration with a test vm, but this apprently expects your data to be available on both ends, and in synch, so basically it expects network based storage which i dont have
[20:11] <patdk-lap2> and the same in reverse at the other side
[20:12] <nacc> tomreyn: did you read `man virsh`? i think you wanted '--copy-storage-all'
[20:12] <tomreyn> well i would still want to encrypt the data in transit
[20:12] <nacc> tomreyn: unless you have base images on both eneds, then you could do --copy-storage-inc, i think
[20:12] <tomreyn> i'll look into '--copy-storage-all', thanks
[20:12] <patdk-lap2> that is going give you a serious limit on your throughput
[20:13] <tomreyn> encryption? yes, it surely will, but i dont mind so much
[20:13] <tomreyn> i can have some downtime, otherwise i'd have started with network based storage in the first place.
[20:13] <nacc> :)
[20:14] <tomreyn> where 'some' is "hours" or, worst case, "a day"
[20:14] <patdk-lap2> that depends on speed, cpu, ...
[20:15] <patdk-lap2> and exactly how large it is
[20:15] <patdk-lap2> could be hours, to 3days for mine atleast
[20:16] <tomreyn> those storages are like 50 GB each, the slowest network link is like 50 mbps, the cpu should be 'good enough' on both ends.
[20:17] <tomreyn> but it's over internet, so weather conditions can change ;)
[20:18] <patdk-lap2> ya, but it depends on if you get multicore support or not
[20:18] <tomreyn> so i guess roughly 2-4 hours per storage
[20:18] <patdk-lap2> actually, that 50mbit will be the limiting factor
[20:18] <tomreyn> i think so, too
[21:47] <an3k> How can I check on which harddrive the linux disk cache is written to and how can I change it?
[21:48] <sarnold> an3k: hmm, this sounds like it needs a lot more context
[21:50] <an3k> Well, I have a DOM on which Ubuntu is installed. Then I additionally have 4 HDDs turned into a RAID10 using mdadm. I often read or write huge files that do not need to stay in disk cache at all.
[21:50] <an3k> However, regardless of that I want to be sure that the DOM is not used for caching anything at all. Everything should be on the RAID
[21:52] <an3k> a) the DOM is very slow and b) it doesn't like plenty of writes because that sooner or later destroys it.
[21:52] <sarnold> an3k: okay; which services would be caching data on the disk?
[21:52] <an3k> par2 and tar but par2 is the main "problem"
[21:53] <sarnold> an3k: presumably you've already made sure you're not swapping to the DOM, right?
[21:53] <an3k> swapoff -a
[21:53] <sarnold> hmm, I don't recall seeing any caching options from tar
[21:54] <an3k> par2 doesn't have any afaik. I hoped I could configure the linux disk cache itself so I don't have to do it for every single application
[21:54] <sarnold> an3k: .. ahhh. I -think- I get it now.
[21:54] <sarnold> an3k: the kernel does not cache disk data onto other disks by default.
[21:54] <sarnold> an3k: you can set up e.g. bcache or ZFS's l2arc if you wish
[21:55] <an3k> So it's caching onto the disk it is installed to?
[21:55] <sarnold> an3k: what tool is telling you about caching?
[21:55] <an3k> free -m
[21:55] <an3k> and htop
[21:55] <sarnold> those are pages in memory, not on block devices
[21:56] <an3k> Hmm, so I misunderstood http://www.linuxatemyram.com/ ? :)
[21:56] <sarnold> I'm afraid so :)
[21:58] <sarnold> the opening sentence is probably where things go awry: "Linux is borrowing unused memory for disk caching"   -- it should probably read "Linux is borrowing unused memory to cache data from the disks."
[21:59] <an3k> so swap is the only "cache on disk" and every other cache is in RAM?
[21:59] <tomreyn> its data previously read from disk, kept there for the purpose of skipping future requests fro the same data to be read from disk again. it is data from disk, cached in ram for faster repeat reads.
[22:00] <an3k> that is awesome but also bad. Bad because I hoped that was the reason (caching on the slow DOM) why the server is kind of slow in the last time ;)
[22:00] <sarnold> an3k: swap is slightly different -- it's memory pages with unique data that hasn't been used in a while. executables (and libraries) can always have that data re-read from disk if needed, but swap is for memory from programs that can't be simply re-created like that. so it gets stored to disk, so more frequently used memory pages can stay in ram.
[22:01] <sarnold> an3k: hehe yeah, I can understand that. finding out if there's actually performance problems, and if so, what you can do about it, is a huge subject..
[22:02] <tomreyn> swap is usually the only cache stored on block devices (this may differ if you are using unusual file systems or services which create their own disk caches etc.)
[22:03] <tomreyn> make sure you are using TRIM on the DOM (if it supports it)
[22:03] <sarnold> hopefully there's not enough writes to the DOM for that to really matter :)
[22:04] <an3k> tomreyn: Nah, that isn't a nice SSD. It's a "crappy" Disk on Module that is slow and cheap but it's tiny and fits directly into a SATA port so ... 1 HE server ;)
[22:04] <an3k> Thank you both for the help and clarification!
[22:05] <tomreyn> welcome ;)
[22:05] <sarnold> an3k: this is probably too in-depth to be useful immediately, but it .. well, it should have more than everything on how the memory system works https://www.kernel.org/doc/gorman/html/understand/
[22:06] <an3k> Holy cow ...
[22:06] <sarnold> and this is a fantastic introduction to one way to figuring out performance bottlenecks http://www.brendangregg.com/usemethod.html
[22:07] <an3k> I didn't even finished the book I bought 6 month ago and that's non-technical and in my native language ... that link will take me ages :)
[22:07] <sarnold> (again, it might be too in-depth but it will help make http://www.brendangregg.com/USEmethod/use-linux.html make sense :)
[22:08] <an3k> I can't give a huge :thumbsup: here but I would if I could ;)
[22:09] <sarnold> :D
[23:33] <JanC> an3k: not sure if that's your problem, but you might also want to put /tmp and other directories that might be used as a "scratchpad" on a faster disk (ramdisk or physical) instead of on the slow one