Re: HAL must die!
On Mar 16, 2011, at 12:29 AM, Matthew Seaman wrote: > > Microsoft may once have had 90% of the desktop market -- but is that > still true? Macs seem to be everywhere nowadays. It may have change a couple of percentage points. Apple marketshare has gone up a lot percentage wise but in the whole market just a little.___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: HAL must die!
On 16/03/2011 00:37, Jerry wrote: > Microsoft has approximately 90% of the desktop market share with > everyone else dividing up the remainder. If you are on a Microsoft > platform you use their products. The same applies to other platforms > and their utilities. Microsoft may once have had 90% of the desktop market -- but is that still true? Macs seem to be everywhere nowadays. Also, how important is 'desktop' nowadays, compared to mobile browsers and the like? If the iPhone doesn't support Flash, then anyone with any sense is going to provide an HTML5 alternative. Cheers, Matthew -- Dr Matthew J Seaman MA, D.Phil. 7 Priory Courtyard Flat 3 PGP: http://www.infracaninophile.co.uk/pgpkey Ramsgate JID: matt...@infracaninophile.co.uk Kent, CT11 9PW signature.asc Description: OpenPGP digital signature
Re: HAL must die!
On 03/16/11 10:43, Chad Perrin wrote: On Tue, Mar 15, 2011 at 08:37:53PM -0400, Jerry wrote: The nice thing about standards is that there are so many of them to choose from. . . . and most of them are supported on any given platform that isn't pathologically closed. Microsoft has approximately 90% of the desktop market share with everyone else dividing up the remainder. If you are on a Microsoft platform you use their products. The same applies to other platforms and their utilities. Corporate media, giants or otherwise, are in business to make money. They obviously are going to focus on the largest possible paying audience. Simple business 101. The largest possible paying audience is generally everybody capable of using an open standard. Thinking that MS Windows users who browse the Web with IE constitute the largest possible paying audience is a classic mistake of not thinking things through. Modern versions (post-6.0) of IE support a nontrivial percentage of open standards; so do Firefox, Opera, Safari, Chromium, and others. If you select standards supported by all of them, you get better than 98% of the user base, and if you select Microsoft technologies, you may get 100% of IE users (post-6.0), but you only get something like 70% of your potential paying audience. People don't target a given proprietary platform that appears to hold the majority of the market because they're targeting the largest possible user base. They do so because they're lazy thinkers. If you hadn't have pointed this out I would have... :) Having studied in some of the best business schools (this means you, Monash, for one) I can vouch for the exceedingly narrow view of the world by the so called captains of industry. Now, as far as HAL goes, the fragmented open-source community cannot even begin to agree on its replacement. Every distro is busy trying to reinvent the wheel. Here you want the majority of users to be dictated to by a minority of users who cannot even agree on a common platform that is uniformly used throughout all the non-Microsoft community. That reasoning is totally irrational. I haven't really been following the goings-on with HAL, so I'm not sure exactly what all is going on there. All I know for sure is that HAL never lived up to its own promises. Clearly, it needed to be either overhauled in a major way or replaced. Just as clearly, there's some kind of confusion over how the solution will look when the dust settles. Beyond that, I'm not sure what's going on. I'm just as mystified. Why do we need a hotplug system on an OS which does it already? It is a system I've never been bothered enough to actually try to understand, its done its job as regards the latest in Xorg (else it wouldn't be needed all), but as for the rest of the features it supposedly does... why? It's easier to read the man pages and research/activate the native functionality in FBSD. I do know, however, that the state of the disunion over HAL is no more or less annoying than the disconnect between API versions in one poorly implemented, incompletely specified, secretly propagated MS Windows version's software framework and another. The real tragedy, I think, is that the majority opinion in any major development space (Apple, Microsoft, Linux, et cetera) is unlikely to be anything clean, elegant, and sane, except in rare cases. Again, my apologies for reviving the dead horse... :) ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: HAL must die!
Erich Dollansky writes: > > The nice thing about standards is that there are so many of them to > > choose from. > > when it comes to screwing, we use - at least outside the USA - > metric screws. M3, M4 ... M10 ... > > We do not care much who manufactured them. > > The software industry is still far away from this. ... in part, because the definition of a "screw" is not yet fixed. Robert Huff ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: looking for rack hosting company
> Sorry if this is a bit off topic. I am looking to run a Freebsd server with a > hosting company. Preferably on the cheaper side as our needs in the beginning > will be somewhat limited. I would need enough ram/storage to do some trial > and > error with different databases. So the vmware virtual servers and limited ram > are out. I recall at one time seeing a very reasonable outfit that would rent > you a rack cpu for around $100/month range and after 2 years would ship it to > you and give you a new one. Can't remember what they were called. I will > consider any suggestions. I already checked out the list of venders at > freebsd.org not finding what I need. Try m5hosting.com . I'm colocate there. They are inexpensive and very good quality hosting company. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: HAL must die!
Hi, On Wednesday 16 March 2011 07:37:53 Jerry wrote: > On Wed, 16 Mar 2011 08:32:01 +1000 > Da Rock articulated: > > > On 03/08/11 03:00, Chad Perrin wrote: > > > On Mon, Mar 07, 2011 at 07:31:43AM -0800, Nerius Landys wrote: > > > > > The nice thing about standards is that there are so many of them to > choose from. when it comes to screwing, we use - at least outside the USA - metric screws. M3, M4 ... M10 ... We do not care much who manufactured them. The software industry is still far away from this. > Now, as far as HAL goes, the fragmented open-source community cannot > even begin to agree on its replacement. Every distro is busy trying to It looks like a bunch of little Napoleons. Erich ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: looking for rack hosting company
> "Darren" == Darren writes: Darren> Sorry if this is a bit off topic. I am looking to run a Freebsd Darren> server with a hosting company. Preferably on the cheaper side Darren> as our needs in the beginning will be somewhat limited. I would Darren> need enough ram/storage to do some trial and error with Darren> different databases. So the vmware virtual servers and limited Darren> ram are out. How so? Take a look at arpnetworks.com/vps - you can get some pretty fat servers there. -- Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 http://www.stonehenge.com/merlyn/> Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc. See http://methodsandmessages.posterous.com/ for Smalltalk discussion ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: HAL must die!
On Tue, 15 Mar 2011, Jerry wrote: Now, as far as HAL goes, the fragmented open-source community cannot even begin to agree on its replacement. Every distro is busy trying to reinvent the wheel. Here you want the majority of users to be dictated to by a minority of users who cannot even agree on a common platform that is uniformly used throughout all the non-Microsoft community. That reasoning is totally irrational. Well you know if HAL were made to work without breaking essential stuff, no one would say a word against it. Unfortunately it was put together by egos too big to make corrections to their beautiful code. -- Lars Eighner http://www.larseighner.com/index.html 8800 N IH35 APT 1191 AUSTIN TX 78753-5266 ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
looking for rack hosting company
Sorry if this is a bit off topic. I am looking to run a Freebsd server with a hosting company. Preferably on the cheaper side as our needs in the beginning will be somewhat limited. I would need enough ram/storage to do some trial and error with different databases. So the vmware virtual servers and limited ram are out. I recall at one time seeing a very reasonable outfit that would rent you a rack cpu for around $100/month range and after 2 years would ship it to you and give you a new one. Can't remember what they were called. I will consider any suggestions. I already checked out the list of venders at freebsd.org not finding what I need. Darren Johnston ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: HAL must die!
On Tue, Mar 15, 2011 at 08:37:53PM -0400, Jerry wrote: > > The nice thing about standards is that there are so many of them to > choose from. . . . and most of them are supported on any given platform that isn't pathologically closed. > > Microsoft has approximately 90% of the desktop market share with > everyone else dividing up the remainder. If you are on a Microsoft > platform you use their products. The same applies to other platforms > and their utilities. > > Corporate media, giants or otherwise, are in business to make money. > They obviously are going to focus on the largest possible paying > audience. Simple business 101. The largest possible paying audience is generally everybody capable of using an open standard. Thinking that MS Windows users who browse the Web with IE constitute the largest possible paying audience is a classic mistake of not thinking things through. Modern versions (post-6.0) of IE support a nontrivial percentage of open standards; so do Firefox, Opera, Safari, Chromium, and others. If you select standards supported by all of them, you get better than 98% of the user base, and if you select Microsoft technologies, you may get 100% of IE users (post-6.0), but you only get something like 70% of your potential paying audience. People don't target a given proprietary platform that appears to hold the majority of the market because they're targeting the largest possible user base. They do so because they're lazy thinkers. > > Now, as far as HAL goes, the fragmented open-source community cannot > even begin to agree on its replacement. Every distro is busy trying to > reinvent the wheel. Here you want the majority of users to be dictated > to by a minority of users who cannot even agree on a common platform > that is uniformly used throughout all the non-Microsoft community. That > reasoning is totally irrational. I haven't really been following the goings-on with HAL, so I'm not sure exactly what all is going on there. All I know for sure is that HAL never lived up to its own promises. Clearly, it needed to be either overhauled in a major way or replaced. Just as clearly, there's some kind of confusion over how the solution will look when the dust settles. Beyond that, I'm not sure what's going on. I do know, however, that the state of the disunion over HAL is no more or less annoying than the disconnect between API versions in one poorly implemented, incompletely specified, secretly propagated MS Windows version's software framework and another. The real tragedy, I think, is that the majority opinion in any major development space (Apple, Microsoft, Linux, et cetera) is unlikely to be anything clean, elegant, and sane, except in rare cases. -- Chad Perrin [ original content licensed OWL: http://owl.apotheon.org ] pgpnSHg4JVwGH.pgp Description: PGP signature
mpd- no ng_l2tp coming up
Been a while since I've had time to check the list- I became a dad again a couple of weeks ago so its been hectic to say the least. Apologies for the previous thread revival- my email client didn't show up with _all_ the emails I missed so I thought I was more recent than I had realised. I'm running into all sorts of issues setting up l2tp networking. I think I have the IPSEC part worked out, but testing parts at a time l2tp dies in a hole. I've resorted to mpd as it seems to be widely used in BSD (and linux too..), but the result seems to be the same for other servers as well. I can start the server, it says ok and runs; I check sockstat, l2tp ports are open; I can even check the console (mpd), its says all systems go. I run the client- the connection dies, and so does the server. mpd.log: Mar 15 23:15:14 bell mpd: Multi-link PPP daemon for FreeBSD Mar 15 23:15:14 bell mpd: Mar 15 23:15:14 bell mpd: process 2762 started, version 5.5 (r...@bell.herveybayaustralia.com.au 10:40 7-Mar-2011) Mar 15 23:15:14 bell mpd: web: listening on 0.0.0.0 5006 Mar 15 23:15:14 bell mpd: EVENT: Registering event EVENT_READ MsgEvent() at msg.c:72 Mar 15 23:15:14 bell mpd: EVENT: Registering event EVENT_READ MsgEvent() done at msg.c:72 Mar 15 23:15:14 bell mpd: EVENT: Registering event EVENT_READ L2tpServerEvent() at l2tp.c:1636 Mar 15 23:15:14 bell mpd: EVENT: Registering event EVENT_READ L2tpServerEvent() done at l2tp.c:1636 Mar 15 23:15:14 bell mpd: L2TP: waiting for connection on 0.0.0.0 1701 Mar 15 23:15:14 bell mpd: EVENT: Processing event EVENT_TIMEOUT ConfigRead() done Mar 15 23:15:36 bell mpd: EVENT: Processing event EVENT_READ L2tpServerEvent() Mar 15 23:15:36 bell mpd: Incoming L2TP packet from 192.168.0.200 47973 Mar 15 23:15:36 bell mpd: L2TP: ppp_l2tp_ctrl_create invoked Mar 15 23:15:36 bell mpd: L2TP: Control connection 0x286f3d08 0.0.0.0 1701 <-> 192.168.0.200 47973 accepted Mar 15 23:15:36 bell mpd: EVENT: Processing event EVENT_READ L2tpServerEvent() done Mar 15 23:15:36 bell mpd: L2TP: RECV [MESSAGE_TYPE SCCRQ] [PROTOCOL_VERSION 1.0] [HOST_NAME "anonymous"] [FRAMING_CAPABILITIES sync=1 async=1] [ASSIGNED_TUNNEL_ID 0x0d78] [RECEIVE_WINDOW_SIZE 1] [CHALLENGE c819a7182517daa2a777da6a7e7e581712745f00e3c707a3f381fb3561faa56e] Mar 15 23:15:36 bell mpd: L2TP: rec'd SCCRQ in state idle Mar 15 23:15:36 bell mpd: L2TP: connected to "anonymous", version=1.0 Mar 15 23:15:36 bell mpd: L2TP: XMIT [MESSAGE_TYPE SCCRP] [HOST_NAME "bell.herveybayaustralia.com.au"] [VENDOR_NAME "FreeBSD MPD"] [BEARER_CAPABILITIES digital=1 analog=1] [RECEIVE_WINDOW_SIZE 8] [PROTOCOL_VERSION 1.0] [FRAMING_CAPABILITIES sync=1 async=1] [ASSIGNED_TUNNEL_ID 0x7008] [CHALLENGE 481df3c95b9e9579adf0cae17d58e680] [CHALLENGE_RESPONSE d6f82bd055e8479f6e8dbe943a5b11c0] Mar 15 23:15:43 bell mpd: L2TP: RECV [MESSAGE_TYPE SCCCN] Mar 15 23:15:43 bell mpd: L2TP: rec'd SCCCN in state wait-ctl-conn Mar 15 23:15:43 bell mpd: L2TP: SCCRP lacks challenge response Mar 15 23:15:43 bell mpd: L2TP: XMIT [MESSAGE_TYPE StopCCN] [ASSIGNED_TUNNEL_ID 0x7008] [RESULT_CODE result=4 error=0 errmsg=""] Mar 15 23:15:43 bell mpd: L2TP: Control connection 0x286f3d08 0.0.0.0 1701 <-> 192.168.0.200 47973 connected Mar 15 23:15:43 bell mpd: L2TP: Control connection 0x286f3d08 terminated: 0 () Mar 15 23:15:43 bell mpd: ASSERT "ctrl->state == CS_DYING" failed: file "l2tp_ctrl.c", line 1426 Mar 15 23:15:43 bell mpd: fatal error, exiting Mar 15 23:15:43 bell mpd: [B1] Bundle: Shutdown Mar 15 23:15:43 bell mpd: [L1] Link: Shutdown Mar 15 23:15:43 bell mpd: L2TP: stop waiting for connection on 0.0.0.0 1701 Mar 15 23:15:43 bell mpd: EVENT: Unregistering event EVENT_READ L2tpServerEvent() at l2tp.c:1671 Mar 15 23:15:43 bell mpd: EVENT: Unregistering event EVENT_READ L2tpServerEvent() done at l2tp.c:1671 Mar 15 23:15:43 bell mpd: PPTP: Total shutdown Mar 15 23:15:43 bell mpd: L2TP: Total shutdown Mar 15 23:15:43 bell mpd: EVENT: Unregistering event EVENT_READ EcpNgDataEvent() at ecp.c:193 Mar 15 23:15:43 bell mpd: EVENT: Unregistering event EVENT_READ EcpNgDataEvent() done at ecp.c:193 Mar 15 23:15:43 bell mpd: EVENT: Unregistering event EVENT_READ CcpNgCtrlEvent() at ccp.c:190 Mar 15 23:15:43 bell mpd: EVENT: Unregistering event EVENT_READ CcpNgCtrlEvent() done at ccp.c:190 Mar 15 23:15:43 bell mpd: EVENT: Unregistering event EVENT_READ CcpNgDataEvent() at ccp.c:193 Mar 15 23:15:43 bell mpd: EVENT: Unregistering event EVENT_READ CcpNgDataEvent() done at ccp.c:193 Mar 15 23:15:43 bell mpd: EVENT: Unregistering event EVENT_READ LinkNgDataEvent() at link.c:182 Mar 15 23:15:43 bell mpd: EVENT: Unregistering event EVENT_READ LinkNgDataEvent() done at link.c:182 Mar 15 23:15:43 bell mpd: process 2762 terminated ngctl list: There are 4 total nodes: Name: mpd4493-cso Type: socket ID: 0048 Num hooks: 0 Name: mpd4493-eso Type: socket ID: 0049 Num hooks: 0 Name: mpd4493-lso Type: socket I
Re: HAL must die!
On Wed, 16 Mar 2011 08:32:01 +1000 Da Rock articulated: > On 03/08/11 03:00, Chad Perrin wrote: > > On Mon, Mar 07, 2011 at 07:31:43AM -0800, Nerius Landys wrote: > > > >> Sorry I could not help but to realize the title of this forum > >> discussion. > >> > >> "I'm sorry Dave, I can't do that." :-/ > >> > > "I'm sorry Dave. I'm afraid I can't do that." > > > > (full quote) > > > > http://www.youtube.com/watch?v=7qnd-hdmgfk > > > > Did you know you can configure YouTube to use HTML5 instead of > > Flash now? Adobe is in danger of becoming irrelevant. > > > Yay! > > Unfortunately that dream may never be satisfied while stupid > corporate media giants have a desire to keep things closed up. And > now we have #%$^%& silverlight to deal with- open source cooperation > my ass! Just got v2 running and now everything requires v3.0... > > That may not affect too much in business, but any corp that has ties > with M$ is pushed to use it., meaning desktop users are left without > entertainment. My wife can't watch some tv sites because of that now- > after finally! getting flash working and watching catch up tv and the > like to come a cropper again with a new format... :( > > If only people would stick to open standards... The nice thing about standards is that there are so many of them to choose from. "Andrew S. Tanenbaum" Microsoft has approximately 90% of the desktop market share with everyone else dividing up the remainder. If you are on a Microsoft platform you use their products. The same applies to other platforms and their utilities. Corporate media, giants or otherwise, are in business to make money. They obviously are going to focus on the largest possible paying audience. Simple business 101. Now, as far as HAL goes, the fragmented open-source community cannot even begin to agree on its replacement. Every distro is busy trying to reinvent the wheel. Here you want the majority of users to be dictated to by a minority of users who cannot even agree on a common platform that is uniformly used throughout all the non-Microsoft community. That reasoning is totally irrational. -- Jerry ✌ freebsd.u...@seibercom.net Disclaimer: off-list followups get on-list replies or get ignored. Please do not ignore the Reply-To header. __ The church is near but the road is icy, the bar is far away but I will walk carefully. Russian Proverb ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: HAL must die!
On 03/08/11 03:00, Chad Perrin wrote: On Mon, Mar 07, 2011 at 07:31:43AM -0800, Nerius Landys wrote: Sorry I could not help but to realize the title of this forum discussion. "I'm sorry Dave, I can't do that." :-/ "I'm sorry Dave. I'm afraid I can't do that." (full quote) http://www.youtube.com/watch?v=7qnd-hdmgfk Did you know you can configure YouTube to use HTML5 instead of Flash now? Adobe is in danger of becoming irrelevant. Yay! Unfortunately that dream may never be satisfied while stupid corporate media giants have a desire to keep things closed up. And now we have #%$^%& silverlight to deal with- open source cooperation my ass! Just got v2 running and now everything requires v3.0... That may not affect too much in business, but any corp that has ties with M$ is pushed to use it., meaning desktop users are left without entertainment. My wife can't watch some tv sites because of that now- after finally! getting flash working and watching catch up tv and the like to come a cropper again with a new format... :( If only people would stick to open standards... ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: can make -j be used for ports?
I had patches for this a couple of years ago when I worked on this problem for summer of code. What I did back then is surely stale, but if people really want it, I'd be happy to take another stab at it. On Tue, Mar 15, 2011 at 3:35 PM, Eitan Adler wrote: >>Is this possible or am I being unreasonable, or both, or not? > > This is unsupported, but you are not being unreasonable. This is a > much wanted feature. > >> Yes. Ports which support parallel builds will have MAKE_JOBS_SAFE=yes set >> in the port Makefile. It defaults to running -j with >> MAKE_JOBS_NUMBER=`${SYSCTL} -n kern.smp.cpus`, but you can change that to >> some other # if you like. > > No, this is incorrect. The MAKE_JOBS_NUMBER and MAKE_JOBS_SAFE is used > internally when building a single port. When the OP is asking if he > can manually specify -j on the command line which would end up > building multiple ports in parallel. This can not be done (primarily > because there is no locking done on ports) > > Certain utilities can make this process faster. For example portmaster > prefetches as much as it can, > > > -- > Eitan Adler > ___ > freebsd-po...@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-ports > To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org" > -- David ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: can make -j be used for ports?
>Is this possible or am I being unreasonable, or both, or not? This is unsupported, but you are not being unreasonable. This is a much wanted feature. > Yes. Ports which support parallel builds will have MAKE_JOBS_SAFE=yes set in > the port Makefile. It defaults to running -j with > MAKE_JOBS_NUMBER=`${SYSCTL} -n kern.smp.cpus`, but you can change that to > some other # if you like. No, this is incorrect. The MAKE_JOBS_NUMBER and MAKE_JOBS_SAFE is used internally when building a single port. When the OP is asking if he can manually specify -j on the command line which would end up building multiple ports in parallel. This can not be done (primarily because there is no locking done on ports) Certain utilities can make this process faster. For example portmaster prefetches as much as it can, -- Eitan Adler ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: IDE -- mount partitions for better performance
On Tue, 15 Mar 2011 17:07:20 -0500, Adam Vande More wrote: > Your statement about master being faster than a slave is simply not true for > almost every scenario when using devices with same capabilites. All > master/slave really controls is enumeration, and shouldn't effect > performance in and of itself. Other variables can effect that of course, > like using a slower device as an ATA Device-1 with a faster Device-0. Even > that example isn't ubiquitous as many, maybe most controllers are able to > support mixed devices each in their fastest mode. My statement originates back from individual experience in settings where disks with different capabilities (esp. very old + very new disk), as well as disk drive and an optical drive with limited speed. > The whole IDE device contention really isn't much of a bottle neck in this > scenario. It's only a big factor when there's *a lot* of simultaneous IO > going to both, say dumping one disk to another. That's true: When copying (or moving) data from one disk to the other master->master seems to be faster than master->slave (same line), if I remember correctly. > The highest preforming setup in something like this is likely to be > something along the lines of a 4-way /boot gmirror, and a 4-way gstripe with > a smaller stripe size eg 32k across the remaining usable space. If you > aggregate your disk IO in this manner, IDE channel contention shouldn't be > much of a bottleneck. A good advice, I haven't thought of that (never tried, but sounds achievable). -- Polytropon Magdeburg, Germany Happy FreeBSD user since 4.0 Andra moi ennepe, Mousa, ... ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: IDE -- mount partitions for better performance
On Tue, Mar 15, 2011 at 3:59 PM, Polytropon wrote: > Keep in mind that performance "across" ad0 and ad2 is best. > Masters are always good. Slaves are slower. Using primary > and secondary in parallel works good, working on a master > and a slave simultanously is worse. > Your statement about master being faster than a slave is simply not true for almost every scenario when using devices with same capabilites. All master/slave really controls is enumeration, and shouldn't effect performance in and of itself. Other variables can effect that of course, like using a slower device as an ATA Device-1 with a faster Device-0. Even that example isn't ubiquitous as many, maybe most controllers are able to support mixed devices each in their fastest mode. The whole IDE device contention really isn't much of a bottle neck in this scenario. It's only a big factor when there's *a lot* of simultaneous IO going to both, say dumping one disk to another. The highest preforming setup in something like this is likely to be something along the lines of a 4-way /boot gmirror, and a 4-way gstripe with a smaller stripe size eg 32k across the remaining usable space. If you aggregate your disk IO in this manner, IDE channel contention shouldn't be much of a bottleneck. -- Adam Vande More ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Dell iDRAC JViewer and FreeBSD 9.0: not working, help!
We use a Dell rack system with a JAVA based virtual console facility, iDRAC6, which works well when used via Windows 7 or some Linux boxes. On FreeBSD 9.0/amd64 (most recent) with JAVA diablo-jdk-1.6.0.07.02_13 Java Development Kit 1.6.0_07.02 openjdk6-b22_2 Oracle's Java 6 virtual machine release under the GPL v2 installed, I receive the virtual console after referring 'javaws' (Java WebStarter) as the appropriate application staring a kind of *.jnl application, but I receive an error: java.lang.reflect.InvocationTargetException at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method) at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39) at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25) at java.lang.reflect.Method.invoke(Method.java:597) at com.sun.javaws.Launcher.executeApplication(Launcher.java:1293) at com.sun.javaws.Launcher.executeMainClass(Launcher.java:1239) at com.sun.javaws.Launcher.doLaunchApp(Launcher.java:1086) at com.sun.javaws.Launcher.run(Launcher.java:105) at java.lang.Thread.run(Thread.java:619) Caused by: java.lang.UnsatisfiedLinkError: com.ami.iusb.FloppyRedir.GetKeyboardName()Ljava/lang/String; at com.ami.iusb.FloppyRedir.GetKeyboardName(Native Method) at com.ami.iusb.FloppyRedir.ReadKeybdType(FloppyRedir.java:454) at com.ami.kvm.jviewer.hid.KVMClient.get_keybd_type(KVMClient.java:731) at com.ami.kvm.jviewer.hid.KVMClient.startRedirection(KVMClient.java:723) at com.ami.kvm.jviewer.gui.JViewerApp.OnVideoStartRedirection(JViewerApp.java:452) at com.ami.kvm.jviewer.gui.JViewerApp.OnConnectToServer(JViewerApp.java:435) at com.ami.kvm.jviewer.JViewer.main(JViewer.java:110) ... 9 more I do not understand much why JAVA, platform independent, is not working with FreeBSD as expected. Linux- und Windows are working. Any help? Oliver ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: IDE -- mount partitions for better performance
On Tue, 15 Mar 2011 15:17:34 -0400, freebsd_u...@guice.ath.cx wrote: > > did I get it right? You have four hard disks? > > Yes, four separate HDD's I assume you have two on each of the two lines: ad0 = primary master ad1 = primary slave ad2 = secondary master ad3 = secondary slave This is how they would be connected: =primary=ad0-ad1 =secondary===ad2-ad3 > Are you suggesting something similar to: > > /dev/ad4s1a for / > /dev/ad4s2a for /tmp > /dev/ad4s3a for /usr > /dev/ad4s4a for /var No. You don't need to even slice the disks (if you're running FreeBSD only, use "dangerously" dedicated layout). But if you require compatibility, make the following layout (just a suggestion): ad0s1a = / ad0s1d = /usr ad1s1a = /tmp ad1s1b = swap ad2s1a = /var ad3s1a = /home Keep in mind that performance "across" ad0 and ad2 is best. Masters are always good. Slaves are slower. Using primary and secondary in parallel works good, working on a master and a slave simultanously is worse. So you have the chance to put different subtrees (for example /usr/obj or /usr/src) on different partitions, drives and lines if needed. The positioning always depends on how much activity you expect on the certain file systems. You did get this idea already. Example: /usr is on prim. master, 2nd partition /usr/obj is mounted to /var/uobj on sec. master, 2nd partition You could also add - with the known limitations (fixed size) - other partitions depending on R/W activity, for example if you need /export or /opt. I have successfully used similar approaches in the past, also using (P)ATA disks, but I had them on a separate controller (also 2 ATA channels, just as you use them from the mainboard). If you omit the slicing step, all examples remove "s1". -- Polytropon Magdeburg, Germany Happy FreeBSD user since 4.0 Andra moi ennepe, Mousa, ... ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: IDE -- mount partitions for better performance
On Tue, Mar 15, 2011 at 3:17 PM, wrote: > Annotated below ... > > > Hi, > > > > On Tuesday 15 March 2011 07:00:30 freebsd_u...@guice.ath.cx wrote: > >> freebsd-questions@freebsd.org > >> > >> Guidance with the following: > >> > >> We are limited to Support for ATA-100/66/33 IDE and ATAPI compliant > >> devices. With that said, we have our atapi/33 optical on a add in > >> controller (PCI) and are seeking to place four HDD’s on the main boards > >> controllers. Our dilemma is where to place /, /tmp, /usr and /var from > >> a > >> performance standpoint. We understand that /var does quite a bit of > >> writing and probably should go on the master hdd, but what about the > >> /usr, > >> /tmp and root? Hell, I’m not sure my thinking is sane as to where I > >> ‘think’ /var should be placed/mounted. > >> > > > > did I get it right? You have four hard disks? > > Yes, four separate HDD's > > > > > If so, place /, /var /tmp on indiidual drives. Make the fourth disk usr > > and mount the remaining space of the other three disks inside /usr/home. > > Are you suggesting something similar to: > > /dev/ad4s1a for / > /dev/ad4s2a for /tmp > /dev/ad4s3a for /usr > /dev/ad4s4a for /var > > If so, my initial can current concern is which device (hdd) from the above > list/configuration, should be connected to which cable connector (master > or slave)? --depending on how much writing to a particular device is > taking place; for instance during a 'build world' or while building > anything from src. there is quite a bit of writing going on. I would > think that making the disk/slice that is being written to a slave would > decrease performance when the master to that slave is also being written > to simultaneously. In such a case the slave would need to wait until the > master is done writing before the slave would be able to write; Is my > thinking on this sane? > > Please enlighten me/us. > > Thank you. > > > > Locate then stuff on the other three disks which you expect to be used in > > parallel with the /usr disk. > > I'm lost on the above suggestion; not understanding this. > > > > Of course, you can mount it anywhere else if you want. > > > > Erich > > > > > > > When four IDE hard disks are usable , then there should be at least two IDE connectors on the main board . IDE cables have two connectors ( connectors near to each other ) . The outer connector is for MASTER , the INNER connector is for SLAVE . IDE hard disks have JUMPERS to set the MASTER and SLAVE hard disks , and it is very likely that on the hard disk , information about jumper settings are printed . When there is no such information printed , it is possible that it is printed in its manual , or it may be available from its producer web site . MASTER hard disk should be connected to outer connector , SLAVE hard disk should be connected to INNER connector . Then , the computer will identify these hard disks with respect to connector on the main board , and MASTER and SLAVE settings of the hard disks . Then , it will list them as /dev/ad1... , /dev/ad2... , /dev/ad3... , /dev/ad4... , or other ad(Numbers) depending on mother board IDE ports settings ( there may be add-on IDE cards ) , or installer may label them differently . If the hard disks are different from each other , it is easy to identify which disk is detected as which device . ( Installer will list information about their sizes ( and , perhaps trade marks ) . ) Assume that all of the hard disks are the same . Then , it is very likely that the connector slot 0 will be detected first , and the connector slot 1 will be detected as second . To be exactly sure , the following steps may be applied . ( Absolutely power down the computer for hardware modifications ) : ( Record detected connected device names . ) ( On the mother board , slot numbers or letters may be printed explicitly . ) Disconnect connector in IDE slot 1 and detect disks in connector in IDE slot 0 . Then , disconnect connector in IDE slot 0 and connect IDE slot 1 connector and detect disks in IDE slot 1. Then , connect connector in IDE slot 0 and detect all of the disks . After obtaining exact device names of the hard disks , final installation may be performed . Thank you very much . Mehmet Erol Sanliturk ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: can make -j be used for ports?
Hi-- [ ...followups sent to freebsd-ports@ list... ] On Mar 15, 2011, at 12:20 PM, John wrote: > I am no expert on make, so, here goes: > > Can -j options be used for make when building ports? Yes. Ports which support parallel builds will have MAKE_JOBS_SAFE=yes set in the port Makefile. It defaults to running -j with MAKE_JOBS_NUMBER=`${SYSCTL} -n kern.smp.cpus`, but you can change that to some other # if you like. Regards, -- -Chuck ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
can make -j be used for ports?
Hello freebsd-questions, I am no expert on make, so, here goes: Can -j options be used for make when building ports? I have an AMD X2 6000+ so presumably I'd use -j3 if this is possible. I know from the docs that -j is not recommended for make installkernel or installworld - but for some ports particularly the metaports I'd like to specify it if only to make the process faster. I'd like to use -j for portsonly and perhaps specify on the commandline when rebuilding the OS from source Is this possible or am I being unreasonable, or both, or not? thanks, -- John - li...@reiteration.net ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: IDE -- mount partitions for better performance
Annotated below ... > Hi, > > On Tuesday 15 March 2011 07:00:30 freebsd_u...@guice.ath.cx wrote: >> freebsd-questions@freebsd.org >> >> Guidance with the following: >> >> We are limited to Support for ATA-100/66/33 IDE and ATAPI compliant >> devices. With that said, we have our atapi/33 optical on a add in >> controller (PCI) and are seeking to place four HDDs on the main boards >> controllers. Our dilemma is where to place /, /tmp, /usr and /var from >> a >> performance standpoint. We understand that /var does quite a bit of >> writing and probably should go on the master hdd, but what about the >> /usr, >> /tmp and root? Hell, Im not sure my thinking is sane as to where I >> think /var should be placed/mounted. >> > did I get it right? You have four hard disks? Yes, four separate HDD's > > If so, place /, /var /tmp on indiidual drives. Make the fourth disk usr > and mount the remaining space of the other three disks inside /usr/home. Are you suggesting something similar to: /dev/ad4s1a for / /dev/ad4s2a for /tmp /dev/ad4s3a for /usr /dev/ad4s4a for /var If so, my initial can current concern is which device (hdd) from the above list/configuration, should be connected to which cable connector (master or slave)? --depending on how much writing to a particular device is taking place; for instance during a 'build world' or while building anything from src. there is quite a bit of writing going on. I would think that making the disk/slice that is being written to a slave would decrease performance when the master to that slave is also being written to simultaneously. In such a case the slave would need to wait until the master is done writing before the slave would be able to write; Is my thinking on this sane? Please enlighten me/us. Thank you. > > Locate then stuff on the other three disks which you expect to be used in > parallel with the /usr disk. I'm lost on the above suggestion; not understanding this. > > Of course, you can mount it anywhere else if you want. > > Erich > > ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: IDE -- mount partitions for better performance
On 14 March 2011 20:00, wrote: > freebsd-questions@freebsd.org > > Guidance with the following: > > We are limited to Support for ATA-100/66/33 IDE and ATAPI compliant > devices. With that said, we have our atapi/33 optical on a add in > controller (PCI) and are seeking to place four HDD’s on the main boards > controllers. Our dilemma is where to place /, /tmp, /usr and /var from a > performance standpoint. We understand that /var does quite a bit of > writing and probably should go on the master hdd, but what about the /usr, > /tmp and root? Hell, I’m not sure my thinking is sane as to where I > ‘think’ /var should be placed/mounted. > It depends on very many things. tmpfs(5) has been extremely stable for me, definitely safe enough for /tmp. Make sure you have plenty of swap, though. If the bits in /var aren't life or death (or at least production mail-server) you can get by with it being mounted on a volatile filesystem (like tmpfs or mdmfs) and backed up occasionally (via cron or similar). If you absolutely can't afford to lose even 5 minutes of /var's past you'd be better off mirroring. I generally put / & /usr on the same disk, they're reasonably small & written to very little, if you symlink /usr/obj, /usr/src, & /usr/ports. -- -- ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"