You may notice squeak.org a bit sluggish over the next few hours. In the past board meeting, we all echoed and reechoed our concern that squeak.org wasn't being backed up anywhere off box. At all. For a very long time. Yes, isn't that crazy? All of squeak.org on a single spindle. Finally, after much coaxing and promising of purchases of fine single-malt scotch (just kidding about that last part, but I wouldn't refuse it if offered), I volunteered some spare disk space and bandwidth on the company server (red.stonehenge.com) to begin a series of steps to modernize our platform. Step 1 is happening now. I'm cloning the image and changes and datafiles that drive squeak.org. I'll then set up a cron job to make sure that is refreshed daily, perhaps even hourly. Step 2 will happen WITH THE HELP OF SOMEONE FROM THE VM TEAM (nudge nudge). I need to build a VM that runs on FreeBSD. If I can do that, I can fire up a "hot spare" site that works like squeak.org although you'll have to get to it with a weird URL. My last attempt to do that failed, and I'm not sure why, except I blamed it on a linux-ism. So if someone FROM THE VM TEAM will please give me precise build instructions for FreeBSD, I'll be happy to give it a try again. Step 3 is very likely (but still in the proposal phase) that the Board will approve me to rent a new VPS (very likely at arpnetworks.com/vps) to build a new home for squeak.org that is on RAID disk and redundant power supply and IPV6 ready and running ZFS for point-in-time snapshots (needed with Squeak for backups). At that point, we'll decommission the old squeak.org, and I'll be the defacto head of the Box Team, if not its only member (since Ken wants to move on, and I don't blame him). So, if squeak.org is a big sluggish while I grab the first 11.2GB snapshot, please be patient. -- Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 <[hidden email]> <URL:http://www.stonehenge.com/merlyn/> Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc. See http://methodsandmessages.posterous.com/ for Smalltalk discussion |
On 18 March 2011 01:52, Randal L. Schwartz <[hidden email]> wrote:
> > You may notice squeak.org a bit sluggish over the next few hours. > > In the past board meeting, we all echoed and reechoed our concern that > squeak.org wasn't being backed up anywhere off box. At all. For a very > long time. Yes, isn't that crazy? All of squeak.org on a single > spindle. > > Finally, after much coaxing and promising of purchases of > fine single-malt scotch (just kidding about that last part, but I > wouldn't refuse it if offered), I volunteered some spare disk space and > bandwidth on the company server (red.stonehenge.com) to begin a series > of steps to modernize our platform. > > Step 1 is happening now. I'm cloning the image and changes and > datafiles that drive squeak.org. I'll then set up a cron job to make > sure that is refreshed daily, perhaps even hourly. > > Step 2 will happen WITH THE HELP OF SOMEONE FROM THE VM TEAM (nudge > nudge). I need to build a VM that runs on FreeBSD. If I can do that, I > can fire up a "hot spare" site that works like squeak.org although > you'll have to get to it with a weird URL. My last attempt to do that > failed, and I'm not sure why, except I blamed it on a linux-ism. So if > someone FROM THE VM TEAM will please give me precise build instructions > for FreeBSD, I'll be happy to give it a try again. > I'd be happy to help and make a build StackVM/Cog configs for FreeBSD. I love this platform , it was my first unix platform i working with. So, if you want to try it out, check the https://pharo-ic.lille.inria.fr/hudson/view/Cog/job/Stack%20VM%20Unix/ there is: https://pharo-ic.lille.inria.fr/hudson/view/Cog/job/Stack%20VM%20Unix/lastSuccessfulBuild/artifact/StackInterpreterUnixConfig-sources.tar.gz which contains all sources ready to build. unpack them, then go to build dir fix paths in directories.cmake file and then run: cmake . make Option #2: gimme shell access to your box, so i can test the config. if we succeed with building stack vm, then we can proceed and try building JIT as well.. But stack VM is simpler for getting started. > Step 3 is very likely (but still in the proposal phase) that the Board > will approve me to rent a new VPS (very likely at arpnetworks.com/vps) > to build a new home for squeak.org that is on RAID disk and redundant > power supply and IPV6 ready and running ZFS for point-in-time snapshots > (needed with Squeak for backups). At that point, we'll decommission the > old squeak.org, and I'll be the defacto head of the Box Team, if not its > only member (since Ken wants to move on, and I don't blame him). > > So, if squeak.org is a big sluggish while I grab the first 11.2GB > snapshot, please be patient. > > -- > Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 > <[hidden email]> <URL:http://www.stonehenge.com/merlyn/> > Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc. > See http://methodsandmessages.posterous.com/ for Smalltalk discussion > > -- Best regards, Igor Stasenko AKA sig. |
In reply to this post by Randal L. Schwartz
On Thu, Mar 17, 2011 at 05:52:46PM -0700, Randal L. Schwartz wrote:
> > You may notice squeak.org a bit sluggish over the next few hours. > > In the past board meeting, we all echoed and reechoed our concern that > squeak.org wasn't being backed up anywhere off box. At all. For a very > long time. Yes, isn't that crazy? All of squeak.org on a single > spindle. > > Finally, after much coaxing and promising of purchases of > fine single-malt scotch (just kidding about that last part, but I > wouldn't refuse it if offered), I volunteered some spare disk space and > bandwidth on the company server (red.stonehenge.com) to begin a series > of steps to modernize our platform. Bravo! Thank you. > > Step 1 is happening now. I'm cloning the image and changes and > datafiles that drive squeak.org. I'll then set up a cron job to make > sure that is refreshed daily, perhaps even hourly. > > Step 2 will happen WITH THE HELP OF SOMEONE FROM THE VM TEAM (nudge > nudge). I need to build a VM that runs on FreeBSD. If I can do that, I > can fire up a "hot spare" site that works like squeak.org although > you'll have to get to it with a weird URL. My last attempt to do that > failed, and I'm not sure why, except I blamed it on a linux-ism. So if > someone FROM THE VM TEAM will please give me precise build instructions > for FreeBSD, I'll be happy to give it a try again. Well, I don't happen to have a FreeBSD box, and my time is going to be really limited over the next week, but what the heck, how hard can it possibly be? ;) So yes I will be *more* than happy to help you with this. Can you give me a pointer to where you left off in the previous attempt (I think I recall it being on one of the mailing lists)? Dave > > Step 3 is very likely (but still in the proposal phase) that the Board > will approve me to rent a new VPS (very likely at arpnetworks.com/vps) > to build a new home for squeak.org that is on RAID disk and redundant > power supply and IPV6 ready and running ZFS for point-in-time snapshots > (needed with Squeak for backups). At that point, we'll decommission the > old squeak.org, and I'll be the defacto head of the Box Team, if not its > only member (since Ken wants to move on, and I don't blame him). > > So, if squeak.org is a big sluggish while I grab the first 11.2GB > snapshot, please be patient. > > -- > Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 > <[hidden email]> <URL:http://www.stonehenge.com/merlyn/> > Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc. > See http://methodsandmessages.posterous.com/ for Smalltalk discussion |
In reply to this post by Randal L. Schwartz
We can all sleep a little better, huge thanks Randal! I don't know
anything about single-malt scotch, but I'll at least buy you a beer if we ever meet. - Chris On Thu, Mar 17, 2011 at 7:52 PM, Randal L. Schwartz <[hidden email]> wrote: > > You may notice squeak.org a bit sluggish over the next few hours. > > In the past board meeting, we all echoed and reechoed our concern that > squeak.org wasn't being backed up anywhere off box. At all. For a very > long time. Yes, isn't that crazy? All of squeak.org on a single > spindle. > > Finally, after much coaxing and promising of purchases of > fine single-malt scotch (just kidding about that last part, but I > wouldn't refuse it if offered), I volunteered some spare disk space and > bandwidth on the company server (red.stonehenge.com) to begin a series > of steps to modernize our platform. > > Step 1 is happening now. I'm cloning the image and changes and > datafiles that drive squeak.org. I'll then set up a cron job to make > sure that is refreshed daily, perhaps even hourly. > > Step 2 will happen WITH THE HELP OF SOMEONE FROM THE VM TEAM (nudge > nudge). I need to build a VM that runs on FreeBSD. If I can do that, I > can fire up a "hot spare" site that works like squeak.org although > you'll have to get to it with a weird URL. My last attempt to do that > failed, and I'm not sure why, except I blamed it on a linux-ism. So if > someone FROM THE VM TEAM will please give me precise build instructions > for FreeBSD, I'll be happy to give it a try again. > > Step 3 is very likely (but still in the proposal phase) that the Board > will approve me to rent a new VPS (very likely at arpnetworks.com/vps) > to build a new home for squeak.org that is on RAID disk and redundant > power supply and IPV6 ready and running ZFS for point-in-time snapshots > (needed with Squeak for backups). At that point, we'll decommission the > old squeak.org, and I'll be the defacto head of the Box Team, if not its > only member (since Ken wants to move on, and I don't blame him). > > So, if squeak.org is a big sluggish while I grab the first 11.2GB > snapshot, please be patient. > > -- > Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 > <[hidden email]> <URL:http://www.stonehenge.com/merlyn/> > Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc. > See http://methodsandmessages.posterous.com/ for Smalltalk discussion > > |
In reply to this post by Igor Stasenko
On Fri, Mar 18, 2011 at 02:01:15AM +0100, Igor Stasenko wrote:
> > On 18 March 2011 01:52, Randal L. Schwartz <[hidden email]> wrote: > > > > You may notice squeak.org a bit sluggish over the next few hours. > > > > In the past board meeting, we all echoed and reechoed our concern that > > squeak.org wasn't being backed up anywhere off box. ??At all. For a very > > long time. ??Yes, isn't that crazy? ??All of squeak.org on a single > > spindle. > > > > Finally, after much coaxing and promising of purchases of > > fine single-malt scotch (just kidding about that last part, but I > > wouldn't refuse it if offered), I volunteered some spare disk space and > > bandwidth on the company server (red.stonehenge.com) to begin a series > > of steps to modernize our platform. > > > > Step 1 is happening now. ??I'm cloning the image and changes and > > datafiles that drive squeak.org. ??I'll then set up a cron job to make > > sure that is refreshed daily, perhaps even hourly. > > > > Step 2 will happen WITH THE HELP OF SOMEONE FROM THE VM TEAM (nudge > > nudge). ??I need to build a VM that runs on FreeBSD. ??If I can do that, I > > can fire up a "hot spare" site that works like squeak.org although > > you'll have to get to it with a weird URL. ??My last attempt to do that > > failed, and I'm not sure why, except I blamed it on a linux-ism. ??So if > > someone FROM THE VM TEAM will please give me precise build instructions > > for FreeBSD, I'll be happy to give it a try again. > > > > I'd be happy to help and make a build StackVM/Cog configs for FreeBSD. > I love this platform , it was my first unix platform i working with. Outstanding, thanks Igor! Randal, my offer to help stands also, but Igor has FreeBSD experience and also can provide a much faster Cog VM, so I would encourage you to take advantage of his offer. You also can accept help from both of us, in which case you will probably end up with more VMs than you know what to do with :) Dave > > So, if you want to try it out, check the > https://pharo-ic.lille.inria.fr/hudson/view/Cog/job/Stack%20VM%20Unix/ > there is: > https://pharo-ic.lille.inria.fr/hudson/view/Cog/job/Stack%20VM%20Unix/lastSuccessfulBuild/artifact/StackInterpreterUnixConfig-sources.tar.gz > which contains all sources ready to build. > unpack them, then go to build dir > fix paths in directories.cmake file > > and then run: > cmake . > make > > Option #2: > > gimme shell access to your box, so i can test the config. > > if we succeed with building stack vm, then we can proceed and try > building JIT as well.. > But stack VM is simpler for getting started. > > > Step 3 is very likely (but still in the proposal phase) that the Board > > will approve me to rent a new VPS (very likely at arpnetworks.com/vps) > > to build a new home for squeak.org that is on RAID disk and redundant > > power supply and IPV6 ready and running ZFS for point-in-time snapshots > > (needed with Squeak for backups). ??At that point, we'll decommission the > > old squeak.org, and I'll be the defacto head of the Box Team, if not its > > only member (since Ken wants to move on, and I don't blame him). > > > > So, if squeak.org is a big sluggish while I grab the first 11.2GB > > snapshot, please be patient. > > > > -- > > Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 > > <[hidden email]> <URL:http://www.stonehenge.com/merlyn/> > > Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc. > > See http://methodsandmessages.posterous.com/ for Smalltalk discussion > > > > > > > > -- > Best regards, > Igor Stasenko AKA sig. |
>>>>> "David" == David T Lewis <[hidden email]> writes:
David> Randal, my offer to help stands also, but Igor has FreeBSD experience David> and also can provide a much faster Cog VM, so I would encourage you to David> take advantage of his offer. You also can accept help from both of us, David> in which case you will probably end up with more VMs than you know what David> to do with :) Thank you! Yes, I'll start on this next week, when the elbow grease refreshes. I've used up my supply tonight. :) -- Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 <[hidden email]> <URL:http://www.stonehenge.com/merlyn/> Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc. See http://methodsandmessages.posterous.com/ for Smalltalk discussion |
Did I miss something? I've been running squeak on freebsd/amd64 for
few years now and I certainly didn't do anything special to build the VM. The stock unix sources compile and run. Didn't try it with Cog yet, but is that needed for squeak.org? Also, doesn't the current machine have disks in raid0? That's the standard config with hetzner. Hetzner also offers some backup space (around 100gigs IIRC, which again, should be enough). Finally, if the Squeak services infrastructure moves to Freebsd I'm willing to donate time to help out with the setup and running it. I have a bunch of servers at Hetzner and very positive experience with them and freebsd - how can linux admins live without jails? I'd argue for renting another machine from them instead of a VPS - why pay more for less? The current Hetzner servers (EQ line) start at quadcore cpu/8gigs of RAM for 49 euro/month. How much would you pay to get that virtualized? Especially the memory tends to be a problem with smalltalks and VPSs. rado On Fri, Mar 18, 2011 at 2:43 AM, Randal L. Schwartz <[hidden email]> wrote: >>>>>> "David" == David T Lewis <[hidden email]> writes: > > David> Randal, my offer to help stands also, but Igor has FreeBSD experience > David> and also can provide a much faster Cog VM, so I would encourage you to > David> take advantage of his offer. You also can accept help from both of us, > David> in which case you will probably end up with more VMs than you know what > David> to do with :) > > Thank you! > > Yes, I'll start on this next week, when the elbow grease refreshes. > I've used up my supply tonight. :) > > -- > Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 > <[hidden email]> <URL:http://www.stonehenge.com/merlyn/> > Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc. > See http://methodsandmessages.posterous.com/ for Smalltalk discussion > > |
>>>>> "radoslav" == radoslav hodnicak <[hidden email]> writes:
radoslav> Finally, if the Squeak services infrastructure moves to Freebsd I'm radoslav> willing to donate time to help out with the setup and running it. I radoslav> have a bunch of servers at Hetzner and very positive experience with radoslav> them and freebsd - how can linux admins live without jails? Thank you! radoslav> I'd argue for renting another machine from them instead of a radoslav> VPS - why pay more for less? The current Hetzner servers (EQ radoslav> line) start at quadcore cpu/8gigs of RAM for 49 radoslav> euro/month. How much would you pay to get that virtualized? radoslav> Especially the memory tends to be a problem with smalltalks radoslav> and VPSs. Take a look at the prices at arpnetworks.com/vps. I'm also able to negotiate some discount over that in exchange for publicity. And I'd rather have a VPS than a hard machine... if one machine goes down, it's up to me to migrate, and immediately. If a VPS host goes down, the hoster has an obligation to launch my VPS on another box. Just one more way to keep a good SLA. -- Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 <[hidden email]> <URL:http://www.stonehenge.com/merlyn/> Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc. See http://methodsandmessages.posterous.com/ for Smalltalk discussion |
On 18.03.2011, at 10:07, Randal L. Schwartz wrote: >>>>>> "radoslav" == radoslav hodnicak <[hidden email]> writes: > > radoslav> Finally, if the Squeak services infrastructure moves to Freebsd I'm > radoslav> willing to donate time to help out with the setup and running it. I > radoslav> have a bunch of servers at Hetzner and very positive experience with > radoslav> them and freebsd - how can linux admins live without jails? > > Thank you! > > radoslav> I'd argue for renting another machine from them instead of a > radoslav> VPS - why pay more for less? The current Hetzner servers (EQ > radoslav> line) start at quadcore cpu/8gigs of RAM for 49 > radoslav> euro/month. How much would you pay to get that virtualized? > radoslav> Especially the memory tends to be a problem with smalltalks > radoslav> and VPSs. > > Take a look at the prices at arpnetworks.com/vps. I'm also able to > negotiate some discount over that in exchange for publicity. And I'd > rather have a VPS than a hard machine... if one machine goes down, it's > up to me to migrate, and immediately. If a VPS host goes down, the > hoster has an obligation to launch my VPS on another box. Just one more > way to keep a good SLA. Do we know how well that virtual host performs compared to a dedicated machine? The services running on squeak.org sometimes are quite busy. One concern with our current Hetzner setup was that the box admins did not have direct contact, but with Marcus as a German-speaking intermediary. Maybe one could switch to direct service in English. In the long run it seems to me we should try to reduce the services we host ourselves. There are always too few people doing sysadmin stuff. In any case, THANKS to Randal for at least doing the backup, and possibly becoming our box admin :) - Bert - |
In reply to this post by radoslav hodnicak
On Fri, 18 Mar 2011, radoslav hodnicak wrote:
> Did I miss something? I've been running squeak on freebsd/amd64 for > few years now and I certainly didn't do anything special to build the > VM. The stock unix sources compile and run. Didn't try it with Cog > yet, but is that needed for squeak.org? Cog would definitely be useful for source.squeak.org. Though I think upgrading to Squeak 4.2 would also improve the performance. > > Also, doesn't the current machine have disks in raid0? That's the Raid0 only gives faster read and write performance. It decreases the reliability of the system. Levente > standard config with hetzner. Hetzner also offers some backup space > (around 100gigs IIRC, which again, should be enough). > > Finally, if the Squeak services infrastructure moves to Freebsd I'm > willing to donate time to help out with the setup and running it. I > have a bunch of servers at Hetzner and very positive experience with > them and freebsd - how can linux admins live without jails? > > I'd argue for renting another machine from them instead of a VPS - why > pay more for less? The current Hetzner servers (EQ line) start at > quadcore cpu/8gigs of RAM for 49 euro/month. How much would you pay to > get that virtualized? Especially the memory tends to be a problem with > smalltalks and VPSs. > > rado > > On Fri, Mar 18, 2011 at 2:43 AM, Randal L. Schwartz > <[hidden email]> wrote: >>>>>>> "David" == David T Lewis <[hidden email]> writes: >> >> David> Randal, my offer to help stands also, but Igor has FreeBSD experience >> David> and also can provide a much faster Cog VM, so I would encourage you to >> David> take advantage of his offer. You also can accept help from both of us, >> David> in which case you will probably end up with more VMs than you know what >> David> to do with :) >> >> Thank you! >> >> Yes, I'll start on this next week, when the elbow grease refreshes. >> I've used up my supply tonight. :) >> >> -- >> Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 >> <[hidden email]> <URL:http://www.stonehenge.com/merlyn/> >> Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc. >> See http://methodsandmessages.posterous.com/ for Smalltalk discussion >> >> > > |
In reply to this post by Randal L. Schwartz
On Thu, 17 Mar 2011, Randal L. Schwartz wrote:
> > You may notice squeak.org a bit sluggish over the next few hours. > > In the past board meeting, we all echoed and reechoed our concern that > squeak.org wasn't being backed up anywhere off box. At all. For a very > long time. Yes, isn't that crazy? All of squeak.org on a single > spindle. > > Finally, after much coaxing and promising of purchases of > fine single-malt scotch (just kidding about that last part, but I > wouldn't refuse it if offered), I volunteered some spare disk space and > bandwidth on the company server (red.stonehenge.com) to begin a series > of steps to modernize our platform. > > Step 1 is happening now. I'm cloning the image and changes and > datafiles that drive squeak.org. I'll then set up a cron job to make > sure that is refreshed daily, perhaps even hourly. Great. > > Step 2 will happen WITH THE HELP OF SOMEONE FROM THE VM TEAM (nudge > nudge). I need to build a VM that runs on FreeBSD. If I can do that, I > can fire up a "hot spare" site that works like squeak.org although > you'll have to get to it with a weird URL. My last attempt to do that > failed, and I'm not sure why, except I blamed it on a linux-ism. So if > someone FROM THE VM TEAM will please give me precise build instructions > for FreeBSD, I'll be happy to give it a try again. Great. > > Step 3 is very likely (but still in the proposal phase) that the Board > will approve me to rent a new VPS (very likely at arpnetworks.com/vps) > to build a new home for squeak.org that is on RAID disk and redundant > power supply and IPV6 ready and running ZFS for point-in-time snapshots > (needed with Squeak for backups). At that point, we'll decommission the > old squeak.org, and I'll be the defacto head of the Box Team, if not its > only member (since Ken wants to move on, and I don't blame him). How much bandwidth and disk does the current server use? Levente > > So, if squeak.org is a big sluggish while I grab the first 11.2GB > snapshot, please be patient. > > -- > Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 > <[hidden email]> <URL:http://www.stonehenge.com/merlyn/> > Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc. > See http://methodsandmessages.posterous.com/ for Smalltalk discussion > > |
In reply to this post by Bert Freudenberg
>>>>> "Bert" == Bert Freudenberg <[hidden email]> writes:
Bert> Do we know how well that virtual host performs compared to a Bert> dedicated machine? The services running on squeak.org sometimes Bert> are quite busy. The VPSs at ARP are quite close to the metal. I'm confident that we could get the performance we need, with the added reduction of risk by having the ability to migrate the entire setup to new hardware at a moment's notice, automatically by the ISP. Bert> One concern with our current Hetzner setup was that the box admins Bert> did not have direct contact, but with Marcus as a German-speaking Bert> intermediary. Maybe one could switch to direct service in English. That would certainly be solved by moving to ARP. Not to mention it being primarily in LA instead of somewhere further away. -- Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 <[hidden email]> <URL:http://www.stonehenge.com/merlyn/> Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc. See http://methodsandmessages.posterous.com/ for Smalltalk discussion |
In reply to this post by Levente Uzonyi-2
>>>>> "Levente" == Levente Uzonyi <[hidden email]> writes:
Levente> How much bandwidth and disk does the current server use? That will be part of my research for step 3. :) By the way, Step 1 is complete! The active part of the disk is now cloned, safely squirreled away in a box in Los Angeles, and backups are being made three times an hour. And, because of ZFS snapshots I'm taking on my machine, we actually have snapshots of *that* at every 10 minutes for two days, every hour for two (four?) weeks, and every day *forever* (or whenever I run out of disk and start stabbing some of the old ones away). ZFS rules. I won't be able to get to step 2 until I get some free time next week. -- Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 <[hidden email]> <URL:http://www.stonehenge.com/merlyn/> Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc. See http://methodsandmessages.posterous.com/ for Smalltalk discussion |
In reply to this post by Randal L. Schwartz
On Fri, Mar 18, 2011 at 10:07 AM, Randal L. Schwartz
<[hidden email]> wrote: > radoslav> I'd argue for renting another machine from them instead of a > radoslav> VPS - why pay more for less? The current Hetzner servers (EQ > radoslav> line) start at quadcore cpu/8gigs of RAM for 49 > radoslav> euro/month. How much would you pay to get that virtualized? > radoslav> Especially the memory tends to be a problem with smalltalks > radoslav> and VPSs. > > Take a look at the prices at arpnetworks.com/vps. I'm also able to > negotiate some discount over that in exchange for publicity. And I'd > rather have a VPS than a hard machine... if one machine goes down, it's > up to me to migrate, and immediately. If a VPS host goes down, the > hoster has an obligation to launch my VPS on another box. Just one more > way to keep a good SLA. I'm looking at the prices. I'd imagine squeak.org needs a machine that has at least 2gigs of ram to run the various squeak images, which is 60 bucks at ARP. 60 bucks at ARP gets you 2gigs RAM 60gigs disk space 600gigs traffic ~65 bucks at Hetzner gets you 8gigs RAM 750gigs disk space 5000gigs traffic I'm not against a virtual machine per se, but those ARP prices don't look appealing to me. And let me repeat that services written in squeak tend to be memory hungry. >> Also, doesn't the current machine have disks in raid0? That's the > > Raid0 only gives faster read and write performance. It decreases the > reliability of the system. Er yeah, I meant mirror. I always get the numbers mixed up rado |
On 3/18/11, radoslav hodnicak <[hidden email]> wrote:
> On Fri, Mar 18, 2011 at 10:07 AM, Randal L. Schwartz > <[hidden email]> wrote: > >> radoslav> I'd argue for renting another machine from them instead of a >> radoslav> VPS - why pay more for less? The current Hetzner servers (EQ >> radoslav> line) start at quadcore cpu/8gigs of RAM for 49 >> radoslav> euro/month. How much would you pay to get that virtualized? >> radoslav> Especially the memory tends to be a problem with smalltalks >> radoslav> and VPSs. >> >> Take a look at the prices at arpnetworks.com/vps. I'm also able to >> negotiate some discount over that in exchange for publicity. And I'd >> rather have a VPS than a hard machine... if one machine goes down, it's >> up to me to migrate, and immediately. If a VPS host goes down, the >> hoster has an obligation to launch my VPS on another box. Just one more >> way to keep a good SLA. > > I'm looking at the prices. I'd imagine squeak.org needs a machine that > has at least 2gigs of ram to run the various squeak images, which is > 60 bucks at ARP. > > 60 bucks at ARP gets you > 2gigs RAM > 60gigs disk space > 600gigs traffic > > ~65 bucks at Hetzner gets you > 8gigs RAM > 750gigs disk space > 5000gigs traffic > > I'm not against a virtual machine per se, but those ARP prices don't > look appealing to me. And let me repeat that services written in > squeak tend to be memory hungry. > >>> Also, doesn't the current machine have disks in raid0? That's the >> >> Raid0 only gives faster read and write performance. It decreases the >> reliability of the system. > > Er yeah, I meant mirror. I always get the numbers mixed up > > rado So you meant Raid 1 --- a server is supposed to have redundancy in the harddisk otherwise it would not be a server. http://en.wikipedia.org/wiki/RAID#Standard_levels RAID 0 (block-level striping without parity or mirroring) has no (or zero) redundancy. "0" NO redundancy In RAID 1 (mirroring without parity or striping), data is written identically to multiple disks (a "mirrored set"). And MOST IMPORTANT http://en.wikipedia.org/wiki/RAID#Data_backup "A RAID system used as a main system disk is not intended as a replacement for backing up data. In parity configurations it will provide a backup-like feature to protect from catastrophic data loss caused by physical damage or errors on a single drive. Many other features of backup systems cannot be provided by RAID arrays alone. The most notable is the ability to restore an earlier version of data, which is needed to protect against software errors causing unwanted data to be written to the disk, and to recover from user error or malicious deletion." Conclusion: So actually a backup should not be too frequent if it only has one version; e.g. a daily or weekly backup would be fine. --Hannes |
In reply to this post by Randal L. Schwartz
On 2011/03/18 15:09, Randal L. Schwartz wrote:
>>>>>> "Levente" == Levente Uzonyi<[hidden email]> writes: > > Levente> How much bandwidth and disk does the current server use? > > That will be part of my research for step 3. :) > > By the way, Step 1 is complete! > > The active part of the disk is now cloned, safely squirreled away in a > box in Los Angeles, and backups are being made three times an hour. > And, because of ZFS snapshots I'm taking on my machine, we actually have > snapshots of *that* at every 10 minutes for two days, every hour for two > (four?) weeks, and every day *forever* (or whenever I run out of disk > and start stabbing some of the old ones away). ZFS rules. It may just be FreeBSD 7 + amd64, but I've a server running in a jail on a machine that USED TO run ZFS. "Used to" because it mostly Just Worked except when it really, really didn't. Apparently the issues have been resolved in FreeBSD 8 though. frank |
In reply to this post by Randal L. Schwartz
Randal wrote:
> If a VPS host goes down, the >hoster has an obligation to launch my VPS on another box. Just one more >way to keep a good SLA. Well, ARP provides unmanaged VPS's for that price. I guess that means you are entitled to a new fresh VPS, which you can then reinstall. As backups are not included, I don't see how they could offer that SLA. That is to say, if I read http://support.arpnetworks.com/kb/vps/what-is-supported correctly. Stephan Eggermont |
On Fri, Mar 18, 2011 at 9:40 AM, Stephan Eggermont <[hidden email]> wrote:
> Randal wrote: >> If a VPS host goes down, the >>hoster has an obligation to launch my VPS on another box. Just one more >>way to keep a good SLA. > > Well, ARP provides unmanaged VPS's for that price. > I guess that means you are entitled to a new fresh VPS, > which you can then reinstall. As backups are not included, > I don't see how they could offer that SLA. You have to make a distinction between the real physical hardware and the virtual machines they host. The physical hardware is the responsibility of ARP. If a power supply blows up, it's their problem and they have to get your virtual machine up and running on another physical machine. OTOH, if you don't pay attention to security and somebody does 'rm -rf /' on your virtual machine, well that's your problem. Randal was talking about the first case, you're talking about the second. > That is to say, if I read > http://support.arpnetworks.com/kb/vps/what-is-supported > correctly. You'll notice that "hardware, and VM hosts" is included in "our responsibility" while "Managing a UNIX-based operating system" is part of "your responsibility". Colin |
Free forum by Nabble | Edit this page |