Samba performance question.
Has anyone had an opportunity to gauge Samba performance on OpenBSD recently, either officially or simply casual observations, I'd love to hear opinions. I'm curious how it might compare to Samba on a modern Linux machine on similar hardware. Thanks in advance.
Re: aterm, rxvt -- memory usage
Jesus Sanchez wrote: Hi all, I'm using 4.2 without problem, and I'm trying to find one "xterm" to my personal use with only one thing in mind: low cpu and memory usage. I have been using mrxvt for years. It's also "multi-tabbed". Currently, I'm running 10 terminals in a single mrxvt process and it is currently using Size: 2644K, and Res: 4472K and barely touches the CPU. I'm running a custom compiled version with all the fancy features turned off. It is extremely fast and reliable; I haven't found anything faster. -pachl
Re: "Correctly" uninstall default Apache and install Apache 2.2.4?
Ed Flecko wrote: Hi folks, For a variety of reasons and features, I'd like to install the apache-httpd-2.2.4.tgz package. As a side note, I tried to install it on OpenBSD 4.2, and there are a few package dependencies it apparently is missing (at least on my box, which runs 4.2 without X) because the install fails. Anyway, 1.) Is there a "correct" way to uninstall the default Apache 1.3 that ships with OpenBSD? I can't use a "pkg_delete..." can I? 2.) Maybe I don't need to? If I don't uninstall the original Apache, will the new version overwrite the 1.3 version? 3.) Do I need to chroot the Apache 2.2.4 or will the "default" install set it up that way? Thank you, Ed I don't think uninstalling the default Apache is recommended or needed. The Apache 2 port should work just fine with the base Apache left alone. Upgrading to 4.3 or a snapshot will get rid of any X dependency problems you may run into as 'needed' cruft like expat is now shipped in base. The Apache 2 port builds the unaltered version from apache.org. I don't believe Apache 2 runs in a chroot by default, so that would be up to you to configure manually. Jason
Re: poll(2) vs kqueue(2) performance
On Mon, Apr 21, 2008 at 09:43:43AM -0500, Marco Peereboom wrote: > On Mon, Apr 21, 2008 at 02:31:26PM +0200, Henning Brauer wrote: > > programming w/ libevent is convenient at times, the decision poll vs > > libevent should not be made based on performance considerations, > > exception beeing the above massive concurrent connection case. > > I spent a lot of time doing libevent stuff for a work project the last > few weeks and I am in love with the API. One negative is the poor > documentation; especially the buffered events docs are inadequate. > > Yes, I read the source :-) but it would have been a little less painful > to have a slightly better man page and some standalone examples. > > Negatives aside once you figure it out it allows one to write code that > doesn't require threading and other complexity inducing stuff. Yay > finite state machines! > Yay, I too fell in love with it and it's various API's despite the lack of documentation for most of them, header help understanding how things work but I wasted quite some time on bufferevents ;-) Gilles -- Gilles Chehade http://www.poolp.org/
Re: QLogic lies: now it's up to you
"Coincidentally", I have just received an update from QLogic in which they emphasize their will to ship us the promised HBAs. Please pause your efforts in sending emails to QLogic for now - it seems it already has helped a lot. I will update you as soon as I hear back from them, which should be end of this week. Thanks guys. On Mon, 2008-04-21 at 11:59 +0200, Stephan A. Rickauer wrote: > In October 2007 I have established contact with QLogic, to investigate > whether they could help us in making iSCSI HBAs work in OpenBSD by > donating some hardware and by providing free programming documentation. > > o;?Unfortunately, Qlogic has chosen to be difficult. This means, I am > forced to make this private communication public to show our user base > the truth about how I and all OpenBSD people have been led along and > lied to by QLogic. Now it's up to us (=you) to show that we won't put up > with that. > > This is the original email conversation between me and QLogic, mainly > Pompey S. Nagra, product manager iSCSI HBAs: > > http://www.ini.uzh.ch/~stephan/qlogic-lies.txt > > After you have read this page, please send an email to the following > people, cite this conversation and request (politely) shipping of the > promised hardware. Do not cite any developer names, QLogic very well > knows who we are and where they should ship the HBAs to (but do cite my > name and the link to the converstation): > > Pompey S. Nagra : [EMAIL PROTECTED] > o;?David Clark: [EMAIL PROTECTED] > Amit Vashi: [EMAIL PROTECTED] > > Please make this issue as public as possible. It would be nice to also > have it on undeadly and other media. > -- Stephan A. Rickauer --- Institute of Neuroinformatics Tel +41 44 635 30 50 University / ETH Zurich Sec +41 44 635 30 52 Winterthurerstrasse 190 Fax +41 44 635 30 53 CH-8057 ZurichWebwww.ini.uzh.ch
Re: "Correctly" uninstall default Apache and install Apache 2.2.4?
Ed Flecko wrote: Hi folks, For a variety of reasons and features, I'd like to install the apache-httpd-2.2.4.tgz package. As a side note, I tried to install it on OpenBSD 4.2, and there are a few package dependencies it apparently is missing (at least on my box, which runs 4.2 without X) because the install fails. http://www.openbsd.org/faq/faq4.html#FilesNeeded http://www.openbsd.org/faq/upgrade42.html#libexpat It was a bug in the 4.2 filesets, expat was moved from the package system to xbase42.tgz, which fewer people install than base42.tgz . 1.) Is there a "correct" way to uninstall the default Apache 1.3 that ships with OpenBSD? I can't use a "pkg_delete..." can I? No. Just leave it. 2.) Maybe I don't need to? If I don't uninstall the original Apache, will the new version overwrite the 1.3 version? If you install the package of Apache 2.2, it won't owerwrite the base Apache. You'll have two Apache installs in two different locations, both of which work and run independently of each other. You may need to double check PATH settings, I'm not sure, but otherwise it should just work if you only run the one you want to run. It's not like the base Apache starts automatically, or anything. 3.) Do I need to chroot the Apache 2.2.4 or will the "default" install set it up that way? I don't have an answer for this one. :-) -- Matthew Weigel hacker [EMAIL PROTECTED]
Re: "Correctly" uninstall default Apache and install Apache 2.2.4?
Ed Flecko schrieb: Hi folks, For a variety of reasons and features, I'd like to install the apache-httpd-2.2.4.tgz package. As a side note, I tried to install it on OpenBSD 4.2, and there are a few package dependencies it apparently is missing (at least on my box, which runs 4.2 without X) because the install fails. Anyway, 1.) Is there a "correct" way to uninstall the default Apache 1.3 that ships with OpenBSD? I can't use a "pkg_delete..." can I? 2.) Maybe I don't need to? If I don't uninstall the original Apache, will the new version overwrite the 1.3 version? 3.) Do I need to chroot the Apache 2.2.4 or will the "default" install set it up that way? Thank you, Ed 1) Not recommended since 1.3 is part of the base distro. Just set httpd_flags="NO" in /etc/rc.conf.local to prevent startup. Dependencies problem is most likely due to expat which is in xbase, you may want to try 4.3 instead which is to be released in a couple of days. 2) No, since packages install somewhere under /usr/local, while base distro is in parent directories. 3) 'find /usr/local/share -type d | grep apache' or similar and look around if you find some useful docs/configuration examples or even startup scripts/snippets that you can use in /etc/rc.local. Regards, Dorian
Re: aterm, rxvt -- memory usage
On Mon, Apr 21 2008 at 34:18, Jesus Sanchez wrote: > Hi all, Hi, > I'm using 4.2 without problem, and I'm trying to find one "xterm" to my > personal use with only one thing in mind: low cpu and memory usage. > > I discarded xterm because it have some things I don't need and it uses a > lot of memory too. > > My two favourite options are aterm and rxvt. I have done some test using > "top" and I have found that aterm uses less memory than rxvt. Where rxvt > uses 1800 KB, aterm uses 1400 KB and stuff like that. I have no problem > using both of them, and I would thing they have the same options for me > except aterm have more package dependences that rxvt, but aterm have > transparency an rxvt not. > > I'm supposed to belive that even having transparency options, aterm uses > less memory than rxvt? I'm missing something? I know they are different > programs but still disturbing me. I personnaly use unicode rxvt. It's a clone of rxvt that comes with unicode (oh surprising) and with client/server mode to reduce memory usage when you have serveral terms like I used to have. urxvt is also one of the rare terms out there with transparency and whitening the background and not darkening it. Claer
Re: "Correctly" uninstall default Apache and install Apache 2.2.4?
On Mon, Apr 21, 2008 at 10:19:18AM -0700, Ed Flecko wrote: | Hi folks, | For a variety of reasons and features, I'd like to install the | apache-httpd-2.2.4.tgz package. As a side note, I tried to install it | on OpenBSD 4.2, and there are a few package dependencies it apparently | is missing (at least on my box, which runs 4.2 without X) because the | install fails. That's interesting. I'm running apache2 on my OpenBSD webserver (which runs 4.2) and it runs just fine there. No problems installing or anything. (although I did have X installed (for gd etc.)) Of course, I installed apache with all its dependencies... | Anyway, | | 1.) Is there a "correct" way to uninstall the default Apache 1.3 that | ships with OpenBSD? I can't use a "pkg_delete..." can I? No. There's no "correct" (as in, documented and advocated by the developers) way of 'uninstalling' the default apache and you can't use pkg_delete (it's not a package in the sense of the pkg-tools). | 2.) Maybe I don't need to? If I don't uninstall the original Apache, | will the new version overwrite the 1.3 version? You don't need to and the 2.2 you want to install will not overwrite the old version. You have to make sure you start the correct apache yourself (it's quite easy to distinguish between the two). | 3.) Do I need to chroot the Apache 2.2.4 or will the "default" install | set it up that way? Not sure about this one, I believe it's not chroot'ed by default but that's easy to check I suppose. Cheers, Paul 'WEiRD' de Weerd -- >[<++>-]<+++.>+++[<-->-]<.>+++[<+ +++>-]<.>++[<>-]<+.--.[-] http://www.weirdnet.nl/
"Correctly" uninstall default Apache and install Apache 2.2.4?
Hi folks, For a variety of reasons and features, I'd like to install the apache-httpd-2.2.4.tgz package. As a side note, I tried to install it on OpenBSD 4.2, and there are a few package dependencies it apparently is missing (at least on my box, which runs 4.2 without X) because the install fails. Anyway, 1.) Is there a "correct" way to uninstall the default Apache 1.3 that ships with OpenBSD? I can't use a "pkg_delete..." can I? 2.) Maybe I don't need to? If I don't uninstall the original Apache, will the new version overwrite the 1.3 version? 3.) Do I need to chroot the Apache 2.2.4 or will the "default" install set it up that way? Thank you, Ed
aterm, rxvt -- memory usage
Hi all, I'm using 4.2 without problem, and I'm trying to find one "xterm" to my personal use with only one thing in mind: low cpu and memory usage. I discarded xterm because it have some things I don't need and it uses a lot of memory too. My two favourite options are aterm and rxvt. I have done some test using "top" and I have found that aterm uses less memory than rxvt. Where rxvt uses 1800 KB, aterm uses 1400 KB and stuff like that. I have no problem using both of them, and I would thing they have the same options for me except aterm have more package dependences that rxvt, but aterm have transparency an rxvt not. I'm supposed to belive that even having transparency options, aterm uses less memory than rxvt? I'm missing something? I know they are different programs but still disturbing me. Thanks for your time.
Re: Question about hard limits
* Thomas Frell <[EMAIL PROTECTED]> [2008-04-21 02:15:29]: > I run an icecast server which may grow into 10 000 listeners or more. > Are their any "gotchas" that I might need to patch change with sysctl? > I've changed kern.maxfiles and wondering if I could run out of TCP > sockets, or file descriptors, or something else I may have missed? > Aiming for high availability, so I am trying to be proactive and find > limits before clients find limits. > > Currently around 4 000 listeners hasn't posed any problems. > > Thanks for any input you may have. > > Sincerely, > > Tom > > > __ > Looking for the perfect gift? Give the gift of Flickr! > > http://www.flickr.com/gift/ > > Also see login.conf(5) and ulimit. -- Travers Buda
Re: can't build ifconfig on landisk
Run "make includes" in /usr/src/include/ first. Damien | hello, | i can't build the ifconfig on landisk. | snapshot is from 2008/04/18 /usr/src is up to date. | | see attached logfile. | | best regards | thomas | cc -O2 -pipe-DINET6 -c ifconfig.c | ifconfig.c: In function `setifwmm': | ifconfig.c:1422: error: storage size of `wmm' isn't known | ifconfig.c:1426: error: `SIOCS80211WMMPARMS' undeclared (first use in this function) | ifconfig.c:1426: error: (Each undeclared identifier is reported only once | ifconfig.c:1426: error: for each function it appears in.) | ifconfig.c: In function `setifwpa': | ifconfig.c:1434: error: storage size of `wpa' isn't known | ifconfig.c:1437: error: `SIOCG80211WPAPARMS' undeclared (first use in this function) | ifconfig.c:1440: error: `SIOCS80211WPAPARMS' undeclared (first use in this function) | ifconfig.c: In function `setifwpaprotos': | ifconfig.c:1448: error: storage size of `wpa' isn't known | ifconfig.c:1457: error: `IEEE80211_WPA_PROTO_WPA1' undeclared (first use in this function) | ifconfig.c:1459: error: `IEEE80211_WPA_PROTO_WPA2' undeclared (first use in this function) | ifconfig.c:1467: error: `SIOCG80211WPAPARMS' undeclared (first use in this function) | ifconfig.c:1470: error: `SIOCS80211WPAPARMS' undeclared (first use in this function) | ifconfig.c: In function `setifwpaakms': | ifconfig.c:1478: error: storage size of `wpa' isn't known | ifconfig.c:1487: error: `IEEE80211_WPA_AKM_PSK' undeclared (first use in this function) | ifconfig.c:1489: error: `IEEE80211_WPA_AKM_IEEE8021X' undeclared (first use in this function) | ifconfig.c:1497: error: `SIOCG80211WPAPARMS' undeclared (first use in this function) | ifconfig.c:1500: error: `SIOCS80211WPAPARMS' undeclared (first use in this function) | ifconfig.c: At top level: | ifconfig.c:1508: error: `IEEE80211_WPA_CIPHER_USEGROUP' undeclared here (not in a function) | ifconfig.c:1508: error: initializer element is not constant | ifconfig.c:1508: error: (near initialization for `ciphers[0].cipher') | ifconfig.c:1508: error: initializer element is not constant | ifconfig.c:1508: error: (near initialization for `ciphers[0]') | ifconfig.c:1509: error: `IEEE80211_WPA_CIPHER_WEP40' undeclared here (not in a function) | ifconfig.c:1509: error: initializer element is not constant | ifconfig.c:1509: error: (near initialization for `ciphers[1].cipher') | ifconfig.c:1509: error: initializer element is not constant | ifconfig.c:1509: error: (near initialization for `ciphers[1]') | ifconfig.c:1510: error: `IEEE80211_WPA_CIPHER_TKIP' undeclared here (not in a function) | ifconfig.c:1510: error: initializer element is not constant | ifconfig.c:1510: error: (near initialization for `ciphers[2].cipher') | ifconfig.c:1510: error: initializer element is not constant | ifconfig.c:1510: error: (near initialization for `ciphers[2]') | ifconfig.c:1511: error: `IEEE80211_WPA_CIPHER_CCMP' undeclared here (not in a function) | ifconfig.c:1511: error: initializer element is not constant | ifconfig.c:1511: error: (near initialization for `ciphers[3].cipher') | ifconfig.c:1511: error: initializer element is not constant | ifconfig.c:1511: error: (near initialization for `ciphers[3]') | ifconfig.c:1512: error: `IEEE80211_WPA_CIPHER_WEP104' undeclared here (not in a function) | ifconfig.c:1512: error: initializer element is not constant | ifconfig.c:1512: error: (near initialization for `ciphers[4].cipher') | ifconfig.c:1512: error: initializer element is not constant | ifconfig.c:1512: error: (near initialization for `ciphers[4]') | ifconfig.c: In function `getwpacipher': | ifconfig.c:1523: error: `IEEE80211_WPA_CIPHER_NONE' undeclared (first use in this function) | ifconfig.c: In function `setifwpaciphers': | ifconfig.c:1530: error: storage size of `wpa' isn't known | ifconfig.c:1539: error: `IEEE80211_WPA_CIPHER_NONE' undeclared (first use in this function) | ifconfig.c:1548: error: `SIOCG80211WPAPARMS' undeclared (first use in this function) | ifconfig.c:1551: error: `SIOCS80211WPAPARMS' undeclared (first use in this function) | ifconfig.c: In function `setifwpagroupcipher': | ifconfig.c:1559: error: storage size of `wpa' isn't known | ifconfig.c:1563: error: `IEEE80211_WPA_CIPHER_NONE' undeclared (first use in this function) | ifconfig.c:1567: error: `SIOCG80211WPAPARMS' undeclared (first use in this function) | ifconfig.c:1570: error: `SIOCS80211WPAPARMS' undeclared (first use in this function) | ifconfig.c: In function `setifwpapsk': | ifconfig.c:1577: error: storage size of `psk' isn't known | ifconfig.c:1592: error: `SIOCS80211WPAPSK' undeclared (first use in this function) | ifconfig.c: In function `print_cipherset': | ifconfig.c:1720: error: `IEEE80211_WPA_CIPHER_NONE' undeclared (first use in this function) | ifconfig.c: In function `ieee80211_status': | ifconfig.c:1739: error: storage size of `psk' isn't known | ifconfig.c:1744: error: storage size of `wmm' isn't known | ifconfig.c:1745: error: storage s
Re: PowerEdge T105
On 2008-04-21, Thomas Martin <[EMAIL PROTECTED]> wrote: > On Fri, Feb 01, 2008 at 08:01:15PM +, Sevan / Venture37 wrote: >> I spent some time today testing Free/Open/NetBSD on the 2 PowerEdges which >> turned up yesterday. >> You can find the dmesgs here: >> http://geeklan.co.uk/files/poweredge_t105/ >> & a brief write up here: >> http://www.geeklan.co.uk/?p=102 > > Hi, > > Has anyone experimented troubles with USB on this machine ? > > I'm unable to mount an FFS-formatted USB key or flash card : the mount > command write an error message "Input/output error" after stalling > during several minutes. I run OpenBSD 4.2. Try a snapshot first...
can't build ifconfig on landisk
hello, i can't build the ifconfig on landisk. snapshot is from 2008/04/18 /usr/src is up to date. see attached logfile. best regards thomas cc -O2 -pipe-DINET6 -c ifconfig.c ifconfig.c: In function `setifwmm': ifconfig.c:1422: error: storage size of `wmm' isn't known ifconfig.c:1426: error: `SIOCS80211WMMPARMS' undeclared (first use in this function) ifconfig.c:1426: error: (Each undeclared identifier is reported only once ifconfig.c:1426: error: for each function it appears in.) ifconfig.c: In function `setifwpa': ifconfig.c:1434: error: storage size of `wpa' isn't known ifconfig.c:1437: error: `SIOCG80211WPAPARMS' undeclared (first use in this function) ifconfig.c:1440: error: `SIOCS80211WPAPARMS' undeclared (first use in this function) ifconfig.c: In function `setifwpaprotos': ifconfig.c:1448: error: storage size of `wpa' isn't known ifconfig.c:1457: error: `IEEE80211_WPA_PROTO_WPA1' undeclared (first use in this function) ifconfig.c:1459: error: `IEEE80211_WPA_PROTO_WPA2' undeclared (first use in this function) ifconfig.c:1467: error: `SIOCG80211WPAPARMS' undeclared (first use in this function) ifconfig.c:1470: error: `SIOCS80211WPAPARMS' undeclared (first use in this function) ifconfig.c: In function `setifwpaakms': ifconfig.c:1478: error: storage size of `wpa' isn't known ifconfig.c:1487: error: `IEEE80211_WPA_AKM_PSK' undeclared (first use in this function) ifconfig.c:1489: error: `IEEE80211_WPA_AKM_IEEE8021X' undeclared (first use in this function) ifconfig.c:1497: error: `SIOCG80211WPAPARMS' undeclared (first use in this function) ifconfig.c:1500: error: `SIOCS80211WPAPARMS' undeclared (first use in this function) ifconfig.c: At top level: ifconfig.c:1508: error: `IEEE80211_WPA_CIPHER_USEGROUP' undeclared here (not in a function) ifconfig.c:1508: error: initializer element is not constant ifconfig.c:1508: error: (near initialization for `ciphers[0].cipher') ifconfig.c:1508: error: initializer element is not constant ifconfig.c:1508: error: (near initialization for `ciphers[0]') ifconfig.c:1509: error: `IEEE80211_WPA_CIPHER_WEP40' undeclared here (not in a function) ifconfig.c:1509: error: initializer element is not constant ifconfig.c:1509: error: (near initialization for `ciphers[1].cipher') ifconfig.c:1509: error: initializer element is not constant ifconfig.c:1509: error: (near initialization for `ciphers[1]') ifconfig.c:1510: error: `IEEE80211_WPA_CIPHER_TKIP' undeclared here (not in a function) ifconfig.c:1510: error: initializer element is not constant ifconfig.c:1510: error: (near initialization for `ciphers[2].cipher') ifconfig.c:1510: error: initializer element is not constant ifconfig.c:1510: error: (near initialization for `ciphers[2]') ifconfig.c:1511: error: `IEEE80211_WPA_CIPHER_CCMP' undeclared here (not in a function) ifconfig.c:1511: error: initializer element is not constant ifconfig.c:1511: error: (near initialization for `ciphers[3].cipher') ifconfig.c:1511: error: initializer element is not constant ifconfig.c:1511: error: (near initialization for `ciphers[3]') ifconfig.c:1512: error: `IEEE80211_WPA_CIPHER_WEP104' undeclared here (not in a function) ifconfig.c:1512: error: initializer element is not constant ifconfig.c:1512: error: (near initialization for `ciphers[4].cipher') ifconfig.c:1512: error: initializer element is not constant ifconfig.c:1512: error: (near initialization for `ciphers[4]') ifconfig.c: In function `getwpacipher': ifconfig.c:1523: error: `IEEE80211_WPA_CIPHER_NONE' undeclared (first use in this function) ifconfig.c: In function `setifwpaciphers': ifconfig.c:1530: error: storage size of `wpa' isn't known ifconfig.c:1539: error: `IEEE80211_WPA_CIPHER_NONE' undeclared (first use in this function) ifconfig.c:1548: error: `SIOCG80211WPAPARMS' undeclared (first use in this function) ifconfig.c:1551: error: `SIOCS80211WPAPARMS' undeclared (first use in this function) ifconfig.c: In function `setifwpagroupcipher': ifconfig.c:1559: error: storage size of `wpa' isn't known ifconfig.c:1563: error: `IEEE80211_WPA_CIPHER_NONE' undeclared (first use in this function) ifconfig.c:1567: error: `SIOCG80211WPAPARMS' undeclared (first use in this function) ifconfig.c:1570: error: `SIOCS80211WPAPARMS' undeclared (first use in this function) ifconfig.c: In function `setifwpapsk': ifconfig.c:1577: error: storage size of `psk' isn't known ifconfig.c:1592: error: `SIOCS80211WPAPSK' undeclared (first use in this function) ifconfig.c: In function `print_cipherset': ifconfig.c:1720: error: `IEEE80211_WPA_CIPHER_NONE' undeclared (first use in this function) ifconfig.c: In function `ieee80211_status': ifconfig.c:1739: error: storage size of `psk' isn't known ifconfig.c:1744: error: storage size of `wmm' isn't known ifconfig.c:1745: error: storage size of `wpa' isn't known ifconfig.c:1763: error: `SIOCG80211WPAPSK' undeclared (first use in this function) ifconfig.c:1783: error: `SIOCG80211WMMPARMS' undeclared (first use in this function) ifconfig.c:178
Re: PowerEdge T105
On Fri, Feb 01, 2008 at 08:01:15PM +, Sevan / Venture37 wrote: > I spent some time today testing Free/Open/NetBSD on the 2 PowerEdges which > turned up yesterday. > You can find the dmesgs here: > http://geeklan.co.uk/files/poweredge_t105/ > & a brief write up here: > http://www.geeklan.co.uk/?p=102 Hi, Has anyone experimented troubles with USB on this machine ? I'm unable to mount an FFS-formatted USB key or flash card : the mount command write an error message "Input/output error" after stalling during several minutes. I run OpenBSD 4.2. /var/log/messages said : Apr 21 15:11:50 winnie2 /bsd: ehci_idone: ex=0xd1210f00 is done! Apr 21 15:12:00 winnie2 /bsd: ehci_sync_hc: tsleep() = 35 Apr 21 15:12:00 winnie2 /bsd: ehci_idone: ex=0xd1210700 is done! Apr 21 15:12:00 winnie2 /bsd: umass0: BBB bulk-out stall clear failed, TIMEOUT Apr 21 15:12:11 winnie2 /bsd: ehci_sync_hc: tsleep() = 35 Apr 21 15:12:11 winnie2 /bsd: ehci_idone: ex=0xd1210a00 is done! Apr 21 15:12:11 winnie2 /bsd: umass0: BBB bulk-out clear stall failed, TIMEOUT Apr 21 15:12:30 winnie2 /bsd: ehci_sync_hc: tsleep() = 35 Apr 21 15:12:30 winnie2 /bsd: ehci_idone: ex=0xd10f9700 is done! Apr 21 15:12:35 winnie2 /bsd: umass0: BBB reset failed, TIMEOUT Apr 21 15:13:12 winnie2 /bsd: ehci_sync_hc: tsleep() = 35 Apr 21 15:13:12 winnie2 /bsd: umass0: BBB reset failed, TIMEOUT Apr 21 15:13:22 winnie2 /bsd: ehci_sync_hc: tsleep() = 35 Apr 21 15:13:22 winnie2 /bsd: ehci_idone: ex=0xd1210a00 is done! Apr 21 15:13:22 winnie2 /bsd: umass0: BBB bulk-out clear stall failed, TIMEOUT -- Thomas Martin <[EMAIL PROTECTED]> GnuPG:1024D/D0AED63E Evolix - Informatique et Logiciels Libres http://www.evolix.fr/ [demime 1.01d removed an attachment of type application/pgp-signature which had a name of signature.asc]
Re: poll(2) vs kqueue(2) performance
On Mon, Apr 21, 2008 at 02:31:26PM +0200, Henning Brauer wrote: > programming w/ libevent is convenient at times, the decision poll vs > libevent should not be made based on performance considerations, > exception beeing the above massive concurrent connection case. I spent a lot of time doing libevent stuff for a work project the last few weeks and I am in love with the API. One negative is the poor documentation; especially the buffered events docs are inadequate. Yes, I read the source :-) but it would have been a little less painful to have a slightly better man page and some standalone examples. Negatives aside once you figure it out it allows one to write code that doesn't require threading and other complexity inducing stuff. Yay finite state machines!
Re: poll(2) vs kqueue(2) performance
Marc, Henning, thank you for the insight. On Mon, Apr 21, 2008 at 8:38 PM, Marc Espie <[EMAIL PROTECTED]> wrote: > > On Sat, Apr 19, 2008 at 11:43:20AM +0200, Jonathan Schleifer wrote: > > "Edwin Eyan Moragas" <[EMAIL PROTECTED]> wrote: > > > > > the question is, which one is more useful when writing new servers? > > > kqueue or poll? > > > > poll is more portable, while kqueue should be more performant (at > > least, that's why it was invented). If your app only needs to run on > > OpenBSD, NetBSD and FreeBSD, you're just fine with kqueue, otherwise > > use poll. Generally, I think it's better to use poll and sacrifice that > > unnoticable performance gain. > > As usual, depends what you want to do. > > poll() and select() give you control over file descriptors. kqueue > encompasses more events. It's not especially faster, it just leads to > simpler code in case you need the supplementary events. > > If all you need to do is watch over a set of file descriptors, poll > and select are the simplest ones to use... and the most portable. > > In many, many cases, poll() is better. The only case where select comes > close is when you want to watch over most of your file descriptors (because > you access less memory in such a case). > > And then, you should profile. I'm not even sure it makes a difference. > > Most of the places in the system where we have select() are legacies: it > it's not broken, don't fix it. > -- garnet:jasmin:beryllium:gluon 90-12264 90-B
Re: Really large drives (was Re: Is there a "badblocks"-equivalent for OpenBSD?)
On Sun, Apr 20, 2008 at 03:35:13PM -0400, Chris Zakelj wrote: > Matthew Weigel wrote: >> Chris Zakelj wrote: >> >>> ... I'm wondering if thought is being given on how to make the physical >>> size (not filesystem... I totally understand why those should be kept >>> small) limitation of http://www.openbsd.org/faq/faq14.html#LargeDrive >> http://www.openbsd.org/43.html >> >> "New Functionality: >> ... >> o The ffs layer is now 64-bit disk block address clean. This means that >> disks, partitions and filesystems larger than 2TB are now supported, with >> the exception of statfs(2) and quotas." >> >> So, yes, thought is being given... > Sweet... I missed that when I did my quick reading of the new features. Is > it safe to assume the guideline of 1M RAM per 1G of file system to do a > reasonable fsck is still valid? It's a bit of an overestimate for the default block and fragment sizes. The main factor for fsck memory needs is the number of inodes in the fileystem being checked. I have a 4TB test filesystem here using the max block and fragment sizes of both 64k. The filesystem has about 17M inodes, and needs about 75M of mem to fsck. Another filesystem (size 48M, using 16k block and 2k fragments) has about 6.5M inodes and needs about 30M of memory. The inode -> mem usage factor is linear: if you double the inodes, you'll need twice the memory. Soon you will hit the maximum data size a process can have. -Otto >> a non-issue on 64-bit platforms >> >> Whether a system is 64-bit or not isn't very relevant to this - that >> mostly establishes what the memory address space is, *not* the size of >> integers that can be used by the system. > Ok... insufficient understanding on my part there :)
Re: short _file in stdio -> fd leak
On Thu, Mar 13, 2008 at 12:56:53PM -0400, Jan Schaumann wrote: > Hey, > > So at work we ran into a situation where a process had to fdopen more > than 32K files, which lead to a file descriptor leak. The reason for > this leak was that while regular fds are ints, _file is a short, so if > fdopen got an fd larger than SHRT_MAX, it would get sign-extended and > thus become invalid, causing the subsequent fclose to fail. > > This being FreeBSD, the fix was found and contributed back into > FreeBSD's repository by John Baldwin in > http://www.freebsd.org/cgi/cvsweb.cgi/src/lib/libc/stdio/fdopen.c.diff?r1=1.8 > ;r2=1.9 > (and other files; with surrounding discussion on > http://docs.freebsd.org/mail/archive/2008/freebsd-arch/20080302.freebsd-arch. > html). > > This fix was also added in NetBSD in > http://mail-index.netbsd.org/source-changes/2008/03/13/msg003463.html . > I believe this also affects OpenBSD and should probably be looked at. > > -Jan > > -- > "When it's fall in New York, the air smells as if someone's been frying > goats in it, and if you are keen to breathe the best plan is to open a > window and stick your head in a building." > > [demime 1.01d removed an attachment of type application/pgp-signature] Thanks for notice. I just committed a fix to our tree. imo the fopen() case is not completely corrrect (_flags is not reset). I sent a note about this to [EMAIL PROTECTED] -Otto
Re: poll(2) vs kqueue(2) performance
On Sat, Apr 19, 2008 at 11:43:20AM +0200, Jonathan Schleifer wrote: > "Edwin Eyan Moragas" <[EMAIL PROTECTED]> wrote: > > > the question is, which one is more useful when writing new servers? > > kqueue or poll? > > poll is more portable, while kqueue should be more performant (at > least, that's why it was invented). If your app only needs to run on > OpenBSD, NetBSD and FreeBSD, you're just fine with kqueue, otherwise > use poll. Generally, I think it's better to use poll and sacrifice that > unnoticable performance gain. As usual, depends what you want to do. poll() and select() give you control over file descriptors. kqueue encompasses more events. It's not especially faster, it just leads to simpler code in case you need the supplementary events. If all you need to do is watch over a set of file descriptors, poll and select are the simplest ones to use... and the most portable. In many, many cases, poll() is better. The only case where select comes close is when you want to watch over most of your file descriptors (because you access less memory in such a case). And then, you should profile. I'm not even sure it makes a difference. Most of the places in the system where we have select() are legacies: it it's not broken, don't fix it.
Re: poll(2) vs kqueue(2) performance
* Edwin Eyan Moragas <[EMAIL PROTECTED]> [2008-04-19 07:34]: > been reading the select(2) man pages and it mentions poll(2) > being more efficient in most cases. this makes it obvious to > discard the use of select(2) in writing new servers. yes. poll is the way better API, easier to use, easier kernel-side, avoids a lot of problems (common!) improper use of select() brings, ... > i've come across some performance benchmarks which is trying > to use kqueue(2). using kqueue directly is painful. I'd recommend libevent if you really want to go that route. That said, kqueue really only pays out when you have _lots_ of concurrent connections, say, >1. > the question is, which one is more useful when writing new servers? > kqueue or poll? programming w/ libevent is convenient at times, the decision poll vs libevent should not be made based on performance considerations, exception beeing the above massive concurrent connection case. -- Henning Brauer, [EMAIL PROTECTED], [EMAIL PROTECTED] BS Web Services, http://bsws.de Full-Service ISP - Secure Hosting, Mail and DNS Services Dedicated Servers, Rootservers, Application Hosting - Hamburg & Amsterdam
Fixed ! (Re: Projector/external monitor not working on OpenBSD 4.2-current on Thinkpad X60)
On Fri, Feb 22, 2008 at 12:05 PM, Matthieu Herrb <[EMAIL PROTECTED]> wrote: > > On Thu, Feb 21, 2008 at 11:11 AM, Amarendra Godbole > <[EMAIL PROTECTED]> wrote: > > I am unable to move the display to a projector or an external monitor > > on my Thinkpad X60, which is running OpenBSD 4.2-current. Fn-F7 is the > > keycombination to be used to switch displays, but it does not work. > > Now, I am not too sure if this is a function of the OS, or Thinkpad's > > firmware. Search engines turned up nothing. Can someone suggest a way > > by which I can make use of an external monitor? Any software package > > to control this? Thanks. > > > > The X60 is using intel i965 graphics right? (hard to tell without some > dmesg or Xorg.0.log attached to your message) > So X is normally using the 'intel' driver which uses XRandR 1.2. > Plug you projector or external monitor, run 'xrandr --auto' and you > should be setup for mirroring. > Check the xrandr(1) man page and the intel web site > http://www.intellinuxgraphics.org/dualhead.html for more configuration > options. Okay, this should have worked earlier (it is i945 chipset), but did not, and I was trying to figure out what must be wrong, until now. I removed my xorg.conf, and then tried doing an "xrandr --auto", and bingo - I had both my LCD and the external monitor working. I am s excited - now I can do all my presentations through OpenBSD, without depending on some other OS to handle it. > > If I wrong and the X60 doesn't use an intel chipset, please post more > details first (Xorg.0.log or dmesg at least) -Amarendra
QLogic lies: now it's up to you
In October 2007 I have established contact with QLogic, to investigate whether they could help us in making iSCSI HBAs work in OpenBSD by donating some hardware and by providing free programming documentation. o;?Unfortunately, Qlogic has chosen to be difficult. This means, I am forced to make this private communication public to show our user base the truth about how I and all OpenBSD people have been led along and lied to by QLogic. Now it's up to us (=you) to show that we won't put up with that. This is the original email conversation between me and QLogic, mainly Pompey S. Nagra, product manager iSCSI HBAs: http://www.ini.uzh.ch/~stephan/qlogic-lies.txt After you have read this page, please send an email to the following people, cite this conversation and request (politely) shipping of the promised hardware. Do not cite any developer names, QLogic very well knows who we are and where they should ship the HBAs to (but do cite my name and the link to the converstation): Pompey S. Nagra : [EMAIL PROTECTED] o;?David Clark: [EMAIL PROTECTED] Amit Vashi: [EMAIL PROTECTED] Please make this issue as public as possible. It would be nice to also have it on undeadly and other media. -- Stephan A. Rickauer --- Institute of Neuroinformatics Tel +41 44 635 30 50 University / ETH Zurich Sec +41 44 635 30 52 Winterthurerstrasse 190 Fax +41 44 635 30 53 CH-8057 ZurichWebwww.ini.uzh.ch
Re: Really large drives (was Re: Is there a "badblocks"-equivalent for OpenBSD?)
On Sun, 2008-04-20 at 22:53 -0500, Matthew Weigel wrote: > David Gwynne wrote: > > > solaris suffers from this problem. you cant use big disks with 32bit > > solaris kernels. > > For UFS, at least, but doesn't ZFS on i386 (not amd64) scale? The filesystem yes, but the block addressing no. I had to split a large raid into lots-of-lessthan1T parts and run zfs on top of them to get a large filesystem on solarisx86-10 lately. [demime 1.01d removed an attachment of type application/pgp-signature which had a name of signature.asc]
Re: OpenBSD 4.2 and spamassasin
Szymon Nowak <[EMAIL PROTECTED]> writes: > Hello I use OpenBSD 4.2 with spamd ( all works ok !! ) > > But a have a question - can I use spamassasin with spamd Yes. Spamassassin also comes with a 'spamd' binary, but it installs in a different location in the file system. The way things usually work, you would set up so spamd (OpenBSD's) helps you throw away the obvious junk, while you let spamassassin examine whatever turns up incoming to your real SMTP server. - P -- Peter N. M. Hansteen, member of the first RFC 1149 implementation team http://bsdly.blogspot.com/ http://www.bsdly.net/ http://www.nuug.no/ "Remember to set the evil bit on all malicious network traffic" delilah spamd[29949]: 85.152.224.147: disconnected after 42673 seconds.
OpenBSD 4.2 and spamassasin
Hello I use OpenBSD 4.2 with spamd ( all works ok !! ) But a have a question - can I use spamassasin with spamd SzymonN [EMAIL PROTECTED] www.0penbsd.pl Visit http://www.inbox.com/email to find out more!
Re: Is there a "badblocks"-equivalent for OpenBSD?
On 2008-04-21, Siegbert Marschall <[EMAIL PROTECTED]> wrote: > i think there are some companies out there having collected a lot > more smart-data the we do, wonder what they do with it... ;) in the case of Google, they wrote a paper, "Failure Trends in a Large Disk Drive Population" (Pinheiro, Weber, Barroso 2007) http://research.google.com/pubs/papers.html
Re: Really large drives (was Re: Is there a "badblocks"-equivalent for OpenBSD?)
On 21/04/2008, at 1:53 PM, Matthew Weigel wrote: David Gwynne wrote: solaris suffers from this problem. you cant use big disks with 32bit solaris kernels. For UFS, at least, but doesn't ZFS on i386 (not amd64) scale? this is a block layer problem, nothing to do with the filesystems. if you boot a 32bit kernel it wont attach driver instances to big disks. anyway, not our problem :) dlg
