Re: pf+FTP external interface only
take a look at : http://mouedine.net/ruleset47.aspx On Thu, 4 Nov 2010 22:27:21 -0700, onteria onte...@scarletdevil.net wrote: I'm currently working on locking down one of my machines with pf. Right now it has a default deny policy and FTP is causing issues. I did a search on how to around FTP oddities using ftp-proxy, but from what I understand this requires an internal interface to work, which this system doesn't have since it's behind a netgear router. Is there something like ftp-proxy for external interface only setups that uses anchors to rewrite rules on the fly? Another option I thought of is making a wrapper script around ftp or whatever the command line client was that would take in the hostname as the first argument, and the rest of the arguments would be passed to whatever the client was. The first call to the script would use pfctl to add the server to a table, which would then have a lenient ruleset for any FTP server in that table. Once the command is done running, pfctl would remove that server from the table. I'm wondering if this would be a good idea. PS: Yes, I plan to setup an OpenBSD router at some point so this doesn't become an issue. Unfortunately I'm saving up for something at the moment, so even a cheap router off Ebay is out of the question right now :) - Onteria
Re: pf+FTP external interface only
On Thu, Nov 4, 2010 at 10:27 PM, onteria onte...@scarletdevil.net wrote: I'm currently working on locking down one of my machines with pf. Right now it has a default deny policy and FTP is causing issues. I did a search on how to around FTP oddities using ftp-proxy, but from what I understand this requires an internal interface to work, which this system doesn't have since it's behind a netgear router. Sounds like your netgear router is handling the NATing and your obsd box is simply a client (single NIC) on the network. Is this correct or am I misreading your description? If correct, you are over-complecating things and do not need ftp-proxy. With pf disabled is FTP working OK? --patrick Is there something like ftp-proxy for external interface only setups that uses anchors to rewrite rules on the fly? Another option I thought of is making a wrapper script around ftp or whatever the command line client was that would take in the hostname as the first argument, and the rest of the arguments would be passed to whatever the client was. The first call to the script would use pfctl to add the server to a table, which would then have a lenient ruleset for any FTP server in that table. Once the command is done running, pfctl would remove that server from the table. I'm wondering if this would be a good idea. PS: Yes, I plan to setup an OpenBSD router at some point so this doesn't become an issue. Unfortunately I'm saving up for something at the moment, so even a cheap router off Ebay is out of the question right now :) - Onteria
Re: OT IPv6 Was: nfsv4?
On 31 October 2010 20:01, Diana Eichert deich...@wrench.com wrote: On Sat, 30 Oct 2010, Marco Peereboom wrote: On Sat, Oct 30, 2010 at 10:02:47AM -0600, Diana Eichert wrote: whether you like it or not, IPv6 deployment is gaining strength. I worked on more exception documents and other excuses than products that would support it ;-) Lets hope the youtubes and facebooks go v6 so that they get of my v4 lawn. excuses only go for so long. I tell you IPv6 deployment is moving forward. think of it as more stimulus money, a lot of h/w will have to be replaced. I just listened to a packetpushers podcast where they specifically mentioned OpenBSD as being one of very few alternatives for ipv6 load balancing. Nice to know that even though it is a cause of unhappiness it is still better supported here. mike
diskmap(4) interface and live USB fstab file
http://www.openbsd.org/faq/faq1.html#WhatsNew says: diskmap(4) interface People using USB attached storage or softraid(4) configurations often had difficulty with drive identifiers changing from boot to boot, or between hardware configurations. diskmap(4) allows you to mount drives by unique disklabel UIDs rather than how they are attached, so now you can use the same /etc/fstab on your USB flash disk without worrying wheter it would come up as sd0, sd1 sd2, etc. http://www.openbsd.org/faq/faq14.html#flashmemLive says: If your target machine has an ahci(4) or SCSI interface, you will probably find your USB drive's identifier changing. Having multiple versions of your /etc/fstab file may make this easier to fix (in single user mode). ---Question Would somebody rewrite #flashmemLive section for the diskmap interface change? or how to edit the /etc/fstab for live USB device without worrying wheter it would come up as sd0, sd1 sd2, etc.
Architeture Choose
Hi All, I'm long time far from OpenBSD world, but planning to come back. The plan is to buy an old machine, but, maybe try an new platform, if the investment worths... I have these options, all in the same price range: A) Sun Fire V100 UltraSPARC IIi 650 Mhz - 2x160Gb Hd - 2Gb RAM - CDROM - US$ 350 B) Apple Power PC G4 733 Mhz - 768 Gb RAM - 38Gb HD - US$ 320,00 C) Atlhon 64 X2 +5200, 2 GB RAM, 160Gb HD - US$ 320,00 The idea is to build an server with: WWW/Email/Firewall funcionalities, with better stablity as possible. I don't think that I will need to upgrade for an period, but pieces that have mechanical components (Hd, cooler) may be a problem, if they are platform-exclusive... Thanks for any help, and sorry for any mistake in my English.. Best Regards, Felipe SP-Brazil
Re: Architeture Choose
On 05-Nov-10 05:47, Felipe Mesquita de Oliveira wrote: C) Atlhon 64 X2 +5200, 2 GB RAM, 160Gb HD - US$ 320,00 The idea is to build an server with: WWW/Email/Firewall funcionalities, with better stablity as possible. You'll get a lot more performance out of the AMD X2. Plus both i386 and AMD64 are still king in the commodity hardware market, and are a dime-a-dozen nowadays. Literally everyone and their grandmothers own x86 based hardware. The i386 platform has support for the most bits of hardware and replacement parts are stupidly easy to come by. -Christopher Ahrens- -Co-founder -LeviaComm Networks-
OpenBSD 4.8: is diskmap(4) missing ?
An earlier post to misc@ made me look into diskmap(4), but the man page seems to be missing: This was a fresh install from CD: # uname -a OpenBSD srv000.home.lan 4.8 GENERIC.MP#335 amd64 # man diskmap man: no entry for diskmap in the manual. # ls /dev/diskmap /dev/diskmap This was an upgrade from CD: gw:remco$ uname -a OpenBSD gw.home.lan 4.8 GENERIC#136 i386 gw:remco$ man diskmap man: no entry for diskmap in the manual. gw:remco$ ls /dev/diskmap /dev/diskmap I don't know if this is the right way to look for the man page, but this comes up empty: gw:OpenBSD$ tar tzf 4.8/i386/man48.tgz |grep diskmap gw:OpenBSD$ tar tzf 4.8/amd64/man48.tgz |grep diskmap gw:OpenBSD$ The on-line manual doesn't really make it clear to me on how to use this either. (http://www.openbsd.org/cgi-bin/man.cgi?query=diskmapsektion=4) The only documentation I was able to find is: http://marc.info/?l=openbsd-miscm=128317640726155w=2
gre mpls packet decapsulation (4.8/i386)
Hello All, a question: i have a setup where an mpls P router sends via GRE SDP (service delivery path) traffic to an openbsd machine, acting as a PE, i need to have my traffic decapsulated to the original payload (minus gre, minus mpls headers) however i have difficulties getting proper payload after a gre interface. following interfaces are configured. original ip addresses are replaced with A and B. vic1: flags=88843UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST,MPLS mtu 1530 lladdr 00:50:56:01:00:9e priority: 0 media: Ethernet autoselect status: active inet ___A___ netmask 0xfff8 broadcast __C__ gre0: flags=89011UP,POINTOPOINT,LINK0,MULTICAST,MPLS mtu 1476 priority: 0 groups: gre physical address inet ___A___-- ___B___ inet ___A___-- ___B___netmask 0xff00 mpe1: flags=51UP,POINTOPOINT,RUNNING mtu 1500 priority: 0 mpls label: 13001 groups: mpe lo1: flags=8049UP,LOOPBACK,RUNNING,MULTICAST mtu 33200 priority: 0 groups: lo inet 192.168.255.1 netmask 0xff00 a packet comes in with following stack, as captured on the vic1. there in MPLS header i have expected label 13001 which should be poped, see attachment, gre-mpls-packet.png, however on the gre0 interface at the same time i see some family 33 header, prepending the payload of original packet starting with the mpls header, see attachment: data-packet.png i run a custom 4.8 kernel on i386 with MP, MPLS and mpe enabled cheers! Vladimir [demime 1.01d removed an attachment of type image/png which had a name of gre-mpls-packet.png] [demime 1.01d removed an attachment of type image/png which had a name of data-packet.png]
Re: gre mpls packet decapsulation (4.8/i386)
On Fri, Nov 05, 2010 at 03:43:07PM +0100, Vladimir Ostrovskiy wrote: Hello All, a question: i have a setup where an mpls P router sends via GRE SDP (service delivery path) traffic to an openbsd machine, acting as a PE, i need to have my traffic decapsulated to the original payload (minus gre, minus mpls headers) however i have difficulties getting proper payload after a gre interface. following interfaces are configured. original ip addresses are replaced with A and B. vic1: flags=88843UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST,MPLS mtu 1530 lladdr 00:50:56:01:00:9e priority: 0 media: Ethernet autoselect status: active inet ___A___ netmask 0xfff8 broadcast __C__ gre0: flags=89011UP,POINTOPOINT,LINK0,MULTICAST,MPLS mtu 1476 priority: 0 groups: gre physical address inet ___A___-- ___B___ inet ___A___-- ___B___netmask 0xff00 mpe1: flags=51UP,POINTOPOINT,RUNNING mtu 1500 priority: 0 mpls label: 13001 groups: mpe lo1: flags=8049UP,LOOPBACK,RUNNING,MULTICAST mtu 33200 priority: 0 groups: lo inet 192.168.255.1 netmask 0xff00 a packet comes in with following stack, as captured on the vic1. there in MPLS header i have expected label 13001 which should be poped, see attachment, gre-mpls-packet.png, however on the gre0 interface at the same time i see some family 33 header, prepending the payload of original packet starting with the mpls header, see attachment: data-packet.png i run a custom 4.8 kernel on i386 with MP, MPLS and mpe enabled Please just include tcpdump -nvi vic1 -X and tcpdump -nvi gre0 -X output. Tcpdump is in base for a reason. Include route -n show -mpls as well please. AF 33 is MPLS and gre(4) so that seems to be OK. -- :wq Claudio
Re: gre mpls packet decapsulation (4.8/i386)
see pcap's attached, On Fri, Nov 5, 2010 at 4:08 PM, Claudio Jeker cje...@diehard.n-r-g.comwrote: On Fri, Nov 05, 2010 at 03:43:07PM +0100, Vladimir Ostrovskiy wrote: Hello All, a question: Please just include tcpdump -nvi vic1 -X and tcpdump -nvi gre0 -X output. Tcpdump is in base for a reason. Include route -n show -mpls as well please. AF 33 is MPLS and gre(4) so that seems to be OK. -- :wq Claudio [demime 1.01d removed an attachment of type application/octet-stream which had a name of gre0-capture.pcap] [demime 1.01d removed an attachment of type application/octet-stream which had a name of vic1-capture.pcap]
Re: gre mpls packet decapsulation (4.8/i386)
forgot the routes # route -n show -mpls Routing tables MPLS: In label Out label Op Gateway Flags Refs Use Mtu Prio Interface 3 - LOCAL default UGT 0 0 - 56 vic0 16 - LOCAL 10.166.41.1 UGT 0 0 - 56 vic0 17 131071 SWAP 10.163.0.161 UGT 0 0 - 56 vic1 18 - LOCAL 192.168.255.1 UGT 0 0 33200 56 lo1 13000 - POP mpe0 T 0 0 - 4 mpe0 13001 - POP mpe1 T 0 14 - 4 mpe1 On Fri, Nov 5, 2010 at 4:08 PM, Claudio Jeker cje...@diehard.n-r-g.comwrote: On Fri, Nov 05, 2010 at 03:43:07PM +0100, Vladimir Ostrovskiy wrote: Hello All, a question: Please just include tcpdump -nvi vic1 -X and tcpdump -nvi gre0 -X output. Tcpdump is in base for a reason. Include route -n show -mpls as well please. AF 33 is MPLS and gre(4) so that seems to be OK. -- :wq Claudio
Re: Architeture Choose
I have an emac that I just updated to 4.8 macppc, and it as expected, it works great.B I used to run OpenBSD on an old ultra5, and it also worked great. x86 might be the most common, but the other architectures work very well too. For what you are doing it looks like all these machines will be fine from a performance standpoint, but as Christopher said, the Athlon will be the snappiest. I'd still get the Sun box though, assuming the fan noise isn't a problem. -- Jeremy Chase http://twitter.com/jeremychase On Fri, Nov 5, 2010 at 9:14 AM, LeviaComm Networks n...@leviacomm.net wrote: On 05-Nov-10 05:47, Felipe Mesquita de Oliveira wrote: C) Atlhon 64 X2 +5200, 2 GB RAM, 160Gb HD - B US$ 320,00 The idea is to build an server with: WWW/Email/Firewall funcionalities, with better stablity as possible. You'll get a lot more performance out of the AMD X2. B Plus both i386 and AMD64 are still king in the commodity hardware market, and are a dime-a-dozen nowadays. B Literally everyone and their grandmothers own x86 based hardware. B The i386 platform has support for the most bits of hardware and replacement parts are stupidly easy to come by. -Christopher Ahrens- -Co-founder -LeviaComm Networks-
pf rules order
Hello there, I posted previously my doubt with the follow subject: 4.7 and ftp-proxy I don't know what are occurring. I have the follow rules: table ftp { address1, address2, address3 } table ftppriv { internal_addr1, internal_addr2 } pass in quick on $int_if proto tcp from ftppriv to port 21 rdr-to 127.0.0.1 port 8021 pass in quick on $int_if proto tcp from $int_inet to ftp port 21 rdr-to 127.0.0.1 port 8021 anchor ftp-proxy/* block log all ... pass in on $int_if proto tcp from 10.1.1.5 From pf.conf man page : For each packet processed by the packet filter, the filter rules are evaluated in sequential order, from first to last. quick If a packet matches a rule which has the quick option set, this rule is considered the last matching rule, and evaluation of sub- sequent rules is skipped. because of this rule pass in on $int_if proto tcp from 10.1.1.5 , this address 10.1.1.5 are accessing every ftp place. If I remove this rule, so it work as expected. Why ? Please can someone explain to me the reason for this? Thanks in advanced
relayd port to linux
Dear Listmember, due to the fact that openssh and some other parts of openbsd are ported to linux maybe you can tell me if you plan to make a openrelayd which is able to compile on linux. I'am willing to try it by my self, maybe you can help me to miss the most common pitfalls ;-). thanks Aleks
Re: relayd port to linux
due to the fact that openssh and some other parts of openbsd are ported to linux maybe you can tell me if you plan to make a openrelayd which is able to compile on linux. relayd depends deeply on pf. so the answer is no.
Re: relayd port to linux
I can only imagine Reyk's face if he saw this. On 11/05/2010 11:54 AM, Aleksandar Lazic wrote: Dear Listmember, due to the fact that openssh and some other parts of openbsd are ported to linux maybe you can tell me if you plan to make a openrelayd which is able to compile on linux. I'am willing to try it by my self, maybe you can help me to miss the most common pitfalls ;-). thanks Aleks -- Joe McDonagh AIM: YoosingYoonickz IRC: joe-mac on freenode When the going gets weird, the weird turn pro.
Re: gre mpls packet decapsulation (4.8/i386)
In gmane.os.openbsd.misc, you wrote: see pcap's attached, On Fri, Nov 5, 2010 at 4:08 PM, Claudio Jeker cje...@diehard.n-r-g.comwrote: On Fri, Nov 05, 2010 at 03:43:07PM +0100, Vladimir Ostrovskiy wrote: Hello All, a question: Please just include tcpdump -nvi vic1 -X and tcpdump -nvi gre0 -X output. Tcpdump is in base for a reason. Include route -n show -mpls as well please. AF 33 is MPLS and gre(4) so that seems to be OK. -- :wq Claudio [demime 1.01d removed an attachment of type application/octet-stream which had a name of gre0-capture.pcap] [demime 1.01d removed an attachment of type application/octet-stream which had a name of vic1-capture.pcap] Just include the output from the commands Claudio showed, pasted into the email body (i.e. in-line text, not as an attachment).
Re: Architeture Choose
On 11/05/10 08:46, Felipe Mesquita de Oliveira wrote: Hi All, I'm long time far from OpenBSD world, but planning to come back. The plan is to buy an old machine, but, maybe try an new platform, if the investment worths... I have these options, all in the same price range: A) Sun Fire V100 UltraSPARC IIi 650 Mhz - 2x160Gb Hd - 2Gb RAM - CDROM - US$ 350 B) Apple Power PC G4 733 Mhz - 768 Gb RAM - 38Gb HD - US$ 320,00 C) Atlhon 64 X2 +5200, 2 GB RAM, 160Gb HD - US$ 320,00 The idea is to build an server with: WWW/Email/Firewall funcionalities, with better stablity as possible. I don't think that I will need to upgrade for an period, but pieces that have mechanical components (Hd, cooler) may be a problem, if they are platform-exclusive... Thanks for any help, and sorry for any mistake in my English.. Best Regards, Felipe SP-Brazil well... Given that choice, I'd go for the Athlon if you need performance (you probably won't), or the Sun Fire v100 if you want to learn something new. I'm not fond of MacPPC machines for the very reason many people love them: the style. The cute cases are a pain in the butt to deal with -- I use a lot of wire rack shelving units, I actually have to velcro-tie the tower macppc systems to the rack to keep the bottom handle from slipping over the front of the shelf and ending up on the floor. The prices on all of them seem high to me, at least in my market. That doesn't mean much. :) One thing to consider is what happens if the box itself fails. OpenBSD is great about moving disks to new hardware in the same platform, but if your Sun fails, you need a compatible sun, if your MacPPC fails, you need another macppc, if your amd64 fails, you need another amd64 (or i386, if you have installed OpenBSD/i386). So, if you run on a macppc or sun system, in the event of failure, you will need to put your hands on a similar machine quickly. The 160G disks in the Sun Fire v100 might hurt you in that regard -- a lot of the Sun IDE disk systems are hw limited to 128G, so you won't be able to stick your 160G disks in an Ultra5, Ultra10, or a Blade100 should your v100 fail. If you go with this machine, I'd put smaller disks in it in case you have to fall back to a U5/U10. If you have to do a cross-platform move, it will require restoring data from your backup, you can't (in general) mount disks from one platform in another and read the data. Nick.
Re: gre mpls packet decapsulation (4.8/i386)
tcpdump -nvi vic1 -X ip proto 47 18:20:27.697032 gre 10.163.0.8 10.163.0.162: [] gre-proto-0x8847 (DF) (ttl 255, id 276, len 130) : 4500 0082 0114 4000 ff2f 6449 0aa3 0008 e.@./dI.#.. 0010: 0aa3 00a2 8847 032c 91ff 0016 4d40 .#G.,...M@ 0020: 17f3 0050 5601 009e 8100 05e6 0800 4500 .s.PV..f..E. 0030: 0054 6891 ff01 3d28 0aa3 00a2 0aa3 .Th....=(.#..# 0040: 0008 0800 17dc c975 4cd4 3cdb 000a .\Iu..LT[.. 0050: a1f1 0809 0a0b 0c0d 0e0f 1011 1213 1415 !q.. 0060: 1617 1819 1a1b .. tcpdump -nvi gre0 -X 18:20:27.697069 MPLS(label 13001, exp 0, ttl 255) : 032c 91ff 0016 4d40 17f3 0050 5601 009e .,....@.s.pv... 0010: 8100 05e6 0800 4500 0054 6891 ff01 ...f..E..Th.... 0020: 3d28 0aa3 00a2 0aa3 0008 0800 17dc c975 =(.#..#.\Iu 0030: 4cd4 3cdb 000a a1f1 0809 0a0b 0c0d ..LT[..!q.. 0040: 0e0f 1011 1213 1415 1617 1819 1a1b 1c1d 0050: 1e1f 2021 2223 2425 2627 2829 2a2b 2c2d .. !#$%'()*+,- 0060: 2e2f 3031 3233 3435 3637 ./01234567 On Fri, Nov 5, 2010 at 6:13 PM, Stuart Henderson s...@spacehopper.orgwrote: In gmane.os.openbsd.misc, you wrote: see pcap's attached, On Fri, Nov 5, 2010 at 4:08 PM, Claudio Jeker cje...@diehard.n-r-g.com wrote: On Fri, Nov 05, 2010 at 03:43:07PM +0100, Vladimir Ostrovskiy wrote: Hello All, a question: Please just include tcpdump -nvi vic1 -X and tcpdump -nvi gre0 -X output. Tcpdump is in base for a reason. Include route -n show -mpls as well please. AF 33 is MPLS and gre(4) so that seems to be OK. -- :wq Claudio [demime 1.01d removed an attachment of type application/octet-stream which had a name of gre0-capture.pcap] [demime 1.01d removed an attachment of type application/octet-stream which had a name of vic1-capture.pcap] Just include the output from the commands Claudio showed, pasted into the email body (i.e. in-line text, not as an attachment).
Re: Architeture Choose
I'm not fond of MacPPC machines for the very reason many people love them: the style. B The cute cases are a pain in the butt to deal with I second that. I had to replace the HD in my emac and I literally had to take the motherboard out to get access. -- Jeremy Chase http://twitter.com/jeremychase On Fri, Nov 5, 2010 at 1:14 PM, Nick Holland n...@holland-consulting.net wrote: On 11/05/10 08:46, Felipe Mesquita de Oliveira wrote: Hi All, I'm long time far from OpenBSD world, but planning to come back. The plan is to buy an old machine, but, maybe try an new platform, if the investment worths... I have these options, all in the same price range: A) Sun Fire V100 UltraSPARC IIi 650 Mhz - 2x160Gb Hd - 2Gb RAM - CDROM - US$ 350 B) Apple Power PC G4 733 Mhz - 768 Gb RAM - 38Gb HD - B US$ 320,00 C) Atlhon 64 X2 +5200, 2 GB RAM, 160Gb HD - B US$ 320,00 The idea is to build an server with: WWW/Email/Firewall funcionalities, with better stablity as possible. I don't think that I will need to upgrade for an period, but pieces that have mechanical components (Hd, cooler) may be a problem, if they are platform-exclusive... Thanks for any help, and sorry for any mistake in my English.. Best Regards, Felipe SP-Brazil well... B Given that choice, I'd go for the Athlon if you need performance (you probably won't), or the Sun Fire v100 if you want to learn something new. I'm not fond of MacPPC machines for the very reason many people love them: the style. B The cute cases are a pain in the butt to deal with -- I use a lot of wire rack shelving units, I actually have to velcro-tie the tower macppc systems to the rack to keep the bottom handle from slipping over the front of the shelf and ending up on the floor. The prices on all of them seem high to me, at least in my market. B That doesn't mean much. B :) One thing to consider is what happens if the box itself fails. B OpenBSD is great about moving disks to new hardware in the same platform, but if your Sun fails, you need a compatible sun, if your MacPPC fails, you need another macppc, if your amd64 fails, you need another amd64 (or i386, if you have installed OpenBSD/i386). B So, if you run on a macppc or sun system, in the event of failure, you will need to put your hands on a similar machine quickly. B The 160G disks in the Sun Fire v100 might hurt you in that regard -- a lot of the Sun IDE disk systems are hw limited to 128G, so you won't be able to stick your 160G disks in an Ultra5, Ultra10, or a Blade100 should your v100 fail. B If you go with this machine, I'd put smaller disks in it in case you have to fall back to a U5/U10. If you have to do a cross-platform move, it will require restoring data from your backup, you can't (in general) mount disks from one platform in another and read the data. Nick.
Re: OpenBSD 4.8 freezes on certain activities
Are you able to try the following? see if it solves your problem. Index: sys/kern/vfs_bio.c === RCS file: /cvs/src/sys/kern/vfs_bio.c,v retrieving revision 1.126 diff -u -r1.126 vfs_bio.c --- sys/kern/vfs_bio.c 3 Aug 2010 06:30:19 - 1.126 +++ sys/kern/vfs_bio.c 5 Nov 2010 17:32:44 - @@ -672,21 +672,10 @@ */ if (!ISSET(bp-b_flags, B_DELWRI)) { SET(bp-b_flags, B_DELWRI); - bp-b_synctime = time_uptime + 35; s = splbio(); reassignbuf(bp); splx(s); curproc-p_stats-p_ru.ru_oublock++;/* XXX */ - } else { - /* -* see if this buffer has slacked through the syncer -* and enforce an async write upon it. -*/ - if (bp-b_synctime time_uptime) { - bawrite(bp); - return; - } - } /* If this is a tape block, write the block now. */ if (major(bp-b_dev) nblkdev @@ -727,7 +716,6 @@ if (ISSET(bp-b_flags, B_DELWRI) == 0) { SET(bp-b_flags, B_DELWRI); - bp-b_synctime = time_uptime + 35; reassignbuf(bp); } } On 3 November 2010 05:17, Michay Koc m...@prime.pl wrote: Hi All, I've just upgraded two of my OpenBSD machines to 4.8: hw.machine=i386 hw.model=Intel(R) Core(TM)2 Duo CPU E8400 @ 3.00GHz (GenuineIntel 686-class) hw.product=DG31PR and hw.machine=i386 hw.model=Intel(R) Atom(TM) CPU D510 @ 1.66GHz (GenuineIntel 686-class) hw.product=D510MO Dmesgs are below. The problem is that they freeze every time I try to: - rsync two local filesystems on different physical disks - high disk IO - about 30GB - run nagios with about 900 probes - hight network IO and ndcpy like 3000 in systat, lots of forks, load average raising to 5 and above High disk IO freeze occurs about 30 seconds after rsync start and is permanent. High network IO freeze occurs several minutes after nagios start and sometimes machines are responsive for limited time. Pkill nagios resolves the problem, machine becomes responsive. In both cases machines behind nat still have internet connectivity. Local services like ssh or console are unavailable. Snapshot from 2010-11-02 22:51:00 does not resolve the issue. The Atom machine freezes much faster than Core2Duo. any help appreciated best regards M.K. Core2Duo dmesg: OpenBSD 4.8 (GENERIC.MP) #359: Mon Aug 16 09:16:26 MDT 2010 dera...@i386.openbsd.org:/usr/src/sys/arch/i386/compile/GENERIC.MP cpu0: Intel(R) Core(TM)2 Duo CPU E8400 @ 3.00GHz (GenuineIntel 686-class) 3.01 GHz cpu0: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUS H,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,VMX,SMX,EST,TM2,S SSE3,CX16,xTPR,PDCM,SSE4.1 real mem = 3476889600 (3315MB) avail mem = 3410038784 (3252MB) mainbus0 at root bios0 at mainbus0: AT/286+ BIOS, date 02/27/08, SMBIOS rev. 2.4 @ 0xe8170 (42 entries) bios0: vendor Intel Corp. version PRG3110H.86A.0047.2008.0227.1745 date 02/27/2008 bios0: Intel Corporation DG31PR acpi0 at bios0: rev 2 acpi0: sleep states S0 S1 S3 S4 S5 acpi0: tables DSDT FACP APIC HPET MCFG acpi0: wakeup devices P0P1(S3) PS2K(S3) PS2M(S3) UAR1(S3) P0P2(S4) USB0(S3) USB1(S3) USB2(S3) USB3(S3) EUSB(S3) MC97(S4) PEX0(S4) PEX1(S4) PEX2(S4) PEX3(S4) SLPB(S4) PWRB(S3) acpitimer0 at acpi0: 3579545 Hz, 24 bits acpimadt0 at acpi0 addr 0xfee0: PC-AT compat cpu0 at mainbus0: apid 0 (boot processor) cpu0: apic clock running at 333MHz cpu1 at mainbus0: apid 1 (application processor) cpu1: Intel(R) Core(TM)2 Duo CPU E8400 @ 3.00GHz (GenuineIntel 686-class) 3 GHz cpu1: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUS H,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,VMX,SMX,EST,TM2,S SSE3,CX16,xTPR,PDCM,SSE4.1 ioapic0 at mainbus0: apid 0 pa 0xfec0, version 20, 24 pins acpihpet0 at acpi0: 14318179 Hz acpiprt0 at acpi0: bus 0 (PCI0) acpiprt1 at acpi0: bus 4 (P0P2) acpiprt2 at acpi0: bus 2 (PEX0) acpiprt3 at acpi0: bus 3 (PEX1) acpiprt4 at acpi0: bus -1 (PEX2) acpiprt5 at acpi0: bus -1 (PEX3) acpicpu0 at acpi0:, C3, C2, C1, PSS acpicpu1 at acpi0:, C3, C2, C1, PSS acpibtn0 at acpi0: SLPB acpibtn1 at acpi0: PWRB bios0: ROM list: 0xc/0xb400! cpu0: Enhanced SpeedStep 3000 MHz: speeds: 2997, 1998 MHz pci0 at mainbus0 bus 0: configuration mode 1 (bios) pchb0 at pci0 dev 0 function 0 Intel 82G33 Host rev 0x10 ppb0 at pci0 dev 1 function 0 Intel 82G33 PCIE rev 0x10: apic 0 int 16 (irq 11) pci1 at ppb0 bus 1 vga1 at pci0 dev 2 function 0 Intel 82G33 Video rev 0x10 wsdisplay0 at vga1 mux 1: console (80x25, vt100 emulation) wsdisplay0: screen 1-5 added (80x25, vt100 emulation) intagp0 at vga1 agp0 at intagp0: aperture at 0xd000, size
Re: Architeture Choose
If your Sun fails -- that's a big IF. It's approaching a possibility of 0 in my experience. If performance isn't an issue and stability is your chief goal, none of this hardware is as stable as a Sun. On 11/05/2010 01:14 PM, Nick Holland wrote: On 11/05/10 08:46, Felipe Mesquita de Oliveira wrote: Hi All, I'm long time far from OpenBSD world, but planning to come back. The plan is to buy an old machine, but, maybe try an new platform, if the investment worths... I have these options, all in the same price range: A) Sun Fire V100 UltraSPARC IIi 650 Mhz - 2x160Gb Hd - 2Gb RAM - CDROM - US$ 350 B) Apple Power PC G4 733 Mhz - 768 Gb RAM - 38Gb HD - US$ 320,00 C) Atlhon 64 X2 +5200, 2 GB RAM, 160Gb HD - US$ 320,00 The idea is to build an server with: WWW/Email/Firewall funcionalities, with better stablity as possible. I don't think that I will need to upgrade for an period, but pieces that have mechanical components (Hd, cooler) may be a problem, if they are platform-exclusive... Thanks for any help, and sorry for any mistake in my English.. Best Regards, Felipe SP-Brazil well... Given that choice, I'd go for the Athlon if you need performance (you probably won't), or the Sun Fire v100 if you want to learn something new. I'm not fond of MacPPC machines for the very reason many people love them: the style. The cute cases are a pain in the butt to deal with -- I use a lot of wire rack shelving units, I actually have to velcro-tie the tower macppc systems to the rack to keep the bottom handle from slipping over the front of the shelf and ending up on the floor. The prices on all of them seem high to me, at least in my market. That doesn't mean much. :) One thing to consider is what happens if the box itself fails. OpenBSD is great about moving disks to new hardware in the same platform, but if your Sun fails, you need a compatible sun, if your MacPPC fails, you need another macppc, if your amd64 fails, you need another amd64 (or i386, if you have installed OpenBSD/i386). So, if you run on a macppc or sun system, in the event of failure, you will need to put your hands on a similar machine quickly. The 160G disks in the Sun Fire v100 might hurt you in that regard -- a lot of the Sun IDE disk systems are hw limited to 128G, so you won't be able to stick your 160G disks in an Ultra5, Ultra10, or a Blade100 should your v100 fail. If you go with this machine, I'd put smaller disks in it in case you have to fall back to a U5/U10. If you have to do a cross-platform move, it will require restoring data from your backup, you can't (in general) mount disks from one platform in another and read the data. Nick. -- Joe McDonagh AIM: YoosingYoonickz IRC: joe-mac on freenode When the going gets weird, the weird turn pro.
Re: Architeture Choose
Back to tha listing =) Thank you everybody for the answers. About the prices, in Brazil we have MercadoLivre (sort of a eBay) Every kind of equipament here is more expensive because of both shipping and fees.. I've translated the prices to US dollar for you to know which choice will be the best cost-benefit option What I really like about the Sun Server was the size... any of the other will take me much more space... BUT, how the guys adviced me, the pictures can't tell how loud the fan can sound... The idea was really to learn something new... I've already used OBSD under i386 with really good results (about a year w/o restart)... I wonder if the other platforms are as good as i386, or even better, form the point of stability... Cheers, Felipe SP-Brazil On Fri, Nov 5, 2010 at 2:25 PM, Jeremy Chase jeremych...@gmail.com wrote: Excellent email, but you didn't send it to the original author. I included him on this forward. :) -- Jeremy Chase http://twitter.com/jeremychase On Fri, Nov 5, 2010 at 1:15 PM, David Astua dast...@gmail.com wrote: 2010/11/5 Jeremy Chase jeremych...@gmail.com: I have an emac that I just updated to 4.8 macppc, and it as expected, it works great.B I used to run OpenBSD on an old ultra5, and it also worked great. x86 might be the most common, but the other architectures work very well too. For what you are doing it looks like all these machines will be fine from a performance standpoint, but as Christopher said, the Athlon will be the snappiest. I'd still get the Sun box though, assuming the fan noise isn't a problem. -- Jeremy Chase http://twitter.com/jeremychase On Fri, Nov 5, 2010 at 9:14 AM, LeviaComm Networks n...@leviacomm.net wrote: On 05-Nov-10 05:47, Felipe Mesquita de Oliveira wrote: C) Atlhon 64 X2 +5200, 2 GB RAM, 160Gb HD - B US$ 320,00 The idea is to build an server with: WWW/Email/Firewall funcionalities, with better stablity as possible. You'll get a lot more performance out of the AMD X2. B Plus both i386 and AMD64 are still king in the commodity hardware market, and are a dime-a-dozen nowadays. B Literally everyone and their grandmothers own x86 based hardware. B The i386 platform has support for the most bits of hardware and replacement parts are stupidly easy to come by. -Christopher Ahrens- -Co-founder -LeviaComm Networks- I've got two old Sun servers one month ago, one of them is a Sunfire like the one you're planing to buy the other is a Netra X1 a bit less powerful. Coincidentally my desktop has the same configuration as the AMD you're mentioning, the performance of the desktop is a bit better, anyway i need to do some further testing. Because think the Sun would respond better under heavy load against the normal performance degradation on my desktop if there's a lot of requests. I'm just messing around with this non-traditional architecture, but take care of the fan noise stated above, the NIC's bundled in the Sun equipments are much better than most on-board NICs, also the LOM interface on the Sun servers is really nice. They're working smoothly! Where are you planning to buy the equipment? I notice that the prices for the equipments are a bit high (for eBay), or you've to pay a lot of shipping/taxes? I hope this helps. Best regards; -- David A. NOTE: If you bought the Sun server don't forget to get the RJ45 - DB9 converter.
Re: Architeture Choose
On Fri, Nov 5, 2010 at 11:30 AM, Joe McDonagh joseph.e.mcdon...@gmail.com wrote: If your Sun fails -- that's a big IF. It's approaching a possibility of 0 in my experience. If performance isn't an issue and stability is your chief goal, none of this hardware is as stable as a Sun. Agreed I've only seen 3 Sun hardware failures (I'm talking about sparcs) in something like 15 years (not counting things like disks or whatever). One was an IPX, that had a motherboard battery die and was easily replaced, but took some work to figure out how to rewrite the prom (after 17 or so years this is still running), another e450 that someone had modified to 'make it faster' and it kept blowing some CPU bridge-thing, and another ultra 1 with an actual logic board failure (it was 10 years old by that point though). as an aside I've thought about putting a bigger disk in the IPX just to see how long it takes to make a release. My netra T1 takes 24 hours and 5.5 seconds to make a full release (including X). Based on absolutely no calculations at all I'd guess a month and 5 seconds. Just for fun: OpenBSD 4.7 (GENERIC) #152: Fri Mar 19 02:33:48 MDT 2010 dera...@sparc.openbsd.org:/usr/src/sys/arch/sparc/compile/GENERIC real mem = 66973696 (63MB) avail mem = 59752448 (56MB) mainbus0 at root: SUNW,Sun 4/50 cpu0 at mainbus0: W8601/8701 or MB86903 @ 40 MHz, on-chip FPU; cache chip bug - trap page uncached cpu0: 64K byte write-through, 32 bytes/line, hw flush cache enabled memreg0 at mainbus0 ioaddr 0xf400 clock0 at mainbus0 ioaddr 0xf200: mk48t02 (eeprom) timer0 at mainbus0 ioaddr 0xf300 delay constant 17 auxreg0 at mainbus0 ioaddr 0xf743 zs0 at mainbus0 ioaddr 0xf100 pri 12, softpri 6 zstty0 at zs0 channel 0 zstty1 at zs0 channel 1 zs1 at mainbus0 ioaddr 0xf000 pri 12, softpri 6 zskbd0 at zs1 channel 0: keyboard, type 5, layout 0x22 wskbd0 at zskbd0: console keyboard zsms0 at zs1 channel 1 wsmouse0 at zsms0 mux 0 audioamd0 at mainbus0 ioaddr 0xf7201000 pri 13, softpri 4 audio0 at audioamd0 sbus0 at mainbus0 ioaddr 0xf800: clock = 20 MHz dma0 at sbus0 slot 0 offset 0x40: rev 1+ esp0 at sbus0 slot 0 offset 0x80 pri 3: ESP100A, 25MHz scsibus0 at esp0: 8 targets, initiator 7 probe(esp0:3:0): max sync rate 8.33MB/s sd0 at scsibus0 targ 3 lun 0: IBMRAID, DFHSS4F9337, 4I4I SCSI2 0/direct fixed sd0: 4303MB, 512 bytes/sec, 8813870 sec total le0 at sbus0 slot 0 offset 0xc0 pri 5: address 08:00:20:08:b4:84 le0: 16 receive buffers, 4 transmit buffers dma1 at sbus0 slot 1 offset 0x81000: rev esc esp1 at dma1 offset 0x8 pri 3: ESP200, 40MHz scsibus1 at esp1: 8 targets, initiator 7 lebuffer0 at sbus0 slot 1 offset 0x4: 128K memory le1 at lebuffer0 offset 0x6 pri 5: address 08:00:20:08:b4:84 le1: 64 receive buffers, 16 transmit buffers dma2 at sbus0 slot 2 offset 0x81000: rev esc esp2 at dma2 offset 0x8 pri 3: ESP200, 25MHz scsibus2 at esp2: 8 targets, initiator 7 lebuffer1 at sbus0 slot 2 offset 0x4: 128K memory le2 at lebuffer1 offset 0x6 pri 5: address 08:00:20:08:b4:84 le2: 64 receive buffers, 16 transmit buffers cgsix0 at sbus0 slot 3 offset 0x0 pri 7: SUNW,501-1672, 1152x900, rev 8 wsdisplay0 at cgsix0 mux 1: console (std, sun emulation), using wskbd0 fdc0 at mainbus0 ioaddr 0xf720 pri 11, softpri 4: chip 82072 fd0 at fdc0 drive 0: 1.44MB 80 cyl, 2 head, 18 sec vscsi0 at root scsibus3 at vscsi0: 256 targets softraid0 at root bootpath: /s...@1,f800/e...@0,80/s...@3,0 root on sd0a swap on sd0b dump on sd0
Re: OpenBSD 4.8 freezes on certain activities
Thank You for your time. The patch seems to resolve both problems on Atom platform. Will check Core2Duo later. Thanks once again Best regard M.K. W dniu 2010-11-05 18:36, Bob Beck pisze: Are you able to try the following? see if it solves your problem. Index: sys/kern/vfs_bio.c === RCS file: /cvs/src/sys/kern/vfs_bio.c,v retrieving revision 1.126 diff -u -r1.126 vfs_bio.c --- sys/kern/vfs_bio.c 3 Aug 2010 06:30:19 - 1.126 +++ sys/kern/vfs_bio.c 5 Nov 2010 17:32:44 - @@ -672,21 +672,10 @@ */ if (!ISSET(bp-b_flags, B_DELWRI)) { SET(bp-b_flags, B_DELWRI); - bp-b_synctime = time_uptime + 35; s = splbio(); reassignbuf(bp); splx(s); curproc-p_stats-p_ru.ru_oublock++;/* XXX */ - } else { - /* -* see if this buffer has slacked through the syncer -* and enforce an async write upon it. -*/ - if (bp-b_synctime time_uptime) { - bawrite(bp); - return; - } - } /* If this is a tape block, write the block now. */ if (major(bp-b_dev) nblkdev @@ -727,7 +716,6 @@ if (ISSET(bp-b_flags, B_DELWRI) == 0) { SET(bp-b_flags, B_DELWRI); - bp-b_synctime = time_uptime + 35; reassignbuf(bp); } } On 3 November 2010 05:17, Michay Kocm...@prime.pl wrote: Hi All, I've just upgraded two of my OpenBSD machines to 4.8: hw.machine=i386 hw.model=Intel(R) Core(TM)2 Duo CPU E8400 @ 3.00GHz (GenuineIntel 686-class) hw.product=DG31PR and hw.machine=i386 hw.model=Intel(R) Atom(TM) CPU D510 @ 1.66GHz (GenuineIntel 686-class) hw.product=D510MO Dmesgs are below. The problem is that they freeze every time I try to: - rsync two local filesystems on different physical disks - high disk IO - about 30GB - run nagios with about 900 probes - hight network IO and ndcpy like 3000 in systat, lots of forks, load average raising to 5 and above High disk IO freeze occurs about 30 seconds after rsync start and is permanent. High network IO freeze occurs several minutes after nagios start and sometimes machines are responsive for limited time. Pkill nagios resolves the problem, machine becomes responsive. In both cases machines behind nat still have internet connectivity. Local services like ssh or console are unavailable. Snapshot from 2010-11-02 22:51:00 does not resolve the issue. The Atom machine freezes much faster than Core2Duo. any help appreciated best regards M.K. Core2Duo dmesg: OpenBSD 4.8 (GENERIC.MP) #359: Mon Aug 16 09:16:26 MDT 2010 dera...@i386.openbsd.org:/usr/src/sys/arch/i386/compile/GENERIC.MP cpu0: Intel(R) Core(TM)2 Duo CPU E8400 @ 3.00GHz (GenuineIntel 686-class) 3.01 GHz cpu0: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,VMX,SMX,EST,TM2,SSSE3,CX16,xTPR,PDCM,SSE4.1 real mem = 3476889600 (3315MB) avail mem = 3410038784 (3252MB) mainbus0 at root bios0 at mainbus0: AT/286+ BIOS, date 02/27/08, SMBIOS rev. 2.4 @ 0xe8170 (42 entries) bios0: vendor Intel Corp. version PRG3110H.86A.0047.2008.0227.1745 date 02/27/2008 bios0: Intel Corporation DG31PR acpi0 at bios0: rev 2 acpi0: sleep states S0 S1 S3 S4 S5 acpi0: tables DSDT FACP APIC HPET MCFG acpi0: wakeup devices P0P1(S3) PS2K(S3) PS2M(S3) UAR1(S3) P0P2(S4) USB0(S3) USB1(S3) USB2(S3) USB3(S3) EUSB(S3) MC97(S4) PEX0(S4) PEX1(S4) PEX2(S4) PEX3(S4) SLPB(S4) PWRB(S3) acpitimer0 at acpi0: 3579545 Hz, 24 bits acpimadt0 at acpi0 addr 0xfee0: PC-AT compat cpu0 at mainbus0: apid 0 (boot processor) cpu0: apic clock running at 333MHz cpu1 at mainbus0: apid 1 (application processor) cpu1: Intel(R) Core(TM)2 Duo CPU E8400 @ 3.00GHz (GenuineIntel 686-class) 3 GHz cpu1: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,VMX,SMX,EST,TM2,SSSE3,CX16,xTPR,PDCM,SSE4.1 ioapic0 at mainbus0: apid 0 pa 0xfec0, version 20, 24 pins acpihpet0 at acpi0: 14318179 Hz acpiprt0 at acpi0: bus 0 (PCI0) acpiprt1 at acpi0: bus 4 (P0P2) acpiprt2 at acpi0: bus 2 (PEX0) acpiprt3 at acpi0: bus 3 (PEX1) acpiprt4 at acpi0: bus -1 (PEX2) acpiprt5 at acpi0: bus -1 (PEX3) acpicpu0 at acpi0:, C3, C2, C1, PSS acpicpu1 at acpi0:, C3, C2, C1, PSS acpibtn0 at acpi0: SLPB acpibtn1 at acpi0: PWRB bios0: ROM list: 0xc/0xb400! cpu0: Enhanced SpeedStep 3000 MHz: speeds: 2997, 1998 MHz pci0 at mainbus0 bus 0: configuration mode 1 (bios) pchb0 at pci0 dev 0 function 0 Intel 82G33 Host rev 0x10 ppb0 at pci0 dev 1 function 0 Intel 82G33 PCIE rev 0x10: apic 0 int 16 (irq 11) pci1 at ppb0 bus 1 vga1 at pci0 dev 2 function 0 Intel 82G33 Video rev 0x10 wsdisplay0 at vga1 mux 1: console (80x25, vt100 emulation)
Le Concours Art7 : plus que 10 jours !!!
ATTENTION !!! Il vous reste plus que 10 jours pour participer au Concours Art7 de Pears Gallery. Voud jtes dij` plus de 300 artistes... Pears Gallery organise du 15 octobre jusqu'au 15 novembre 2010 un concours riservi aux artistes professionnels frangais (catigories : peintures, estampes, photographies d'art et art numirique). Le jury de Pears Gallery silectionnera 7 oeuvres ( les 3 coups de coeur du jury et une oeuvre par catigorie). Les artistes ricompensis gagneront 3 mois d'exposition gratuite sur le site www.pears-gallery.com (7.000 ` 13.000 visiteurs chaque mois) et les oeuvres seront prisenties dans la Newsletter du site (diffusion ` plus de 10.000 abonnis). La participation ` ce concours est gratuite. Comment participer ? 1. Vous devez vous inscrire (gratuit) sur notre site : http://www.pears-gallery.com/fr/register/register1 2. Un mail de confirmation vous sera envoyi avec un lien de validation. 3. En confirmant votre inscription, vous arrivez dans votre espace privi sur une page tarifs/abonnement. VOUS N'ETES PAS OBLIGE DE VOUS ABONNER. 3. Dans votre espace privi, en cliquant sur l'onglet OEUVRE, vous insirez vos oeuvres (descriptif + photos), nous vous recommandons d'utiliser toutes les fonctions du site et d'y ajouter des photos de ditails. A bienttt. L'Equipe Pears Gallery Disinscription
Re: OpenBSD 4.8 freezes on certain activities
Hmmm, I was a little bit too optimistic. The hight disk IO seems not to cause problems now, but network io (re adapter) from nagios(probably) has freezed the Atom machine after approximately 2 hours. This is top header right after freeze: 75 processes: 1 running, 70 idle, 4 on processor CPU0 states: 2.9% user, 0.0% nice, 17.3% system, 51.9% interrupt, 27.9% idle CPU1 states: 4.6% user, 0.0% nice, 80.2% system, 4.6% interrupt, 10.7% idle CPU2 states: 25.2% user, 0.0% nice, 54.8% system, 17.0% interrupt, 3.0% idle CPU3 states: 8.0% user, 0.0% nice, 52.1% system, 18.4% interrupt, 21.5% idle Memory: Real: 106M/453M act/tot Free: 2794M Swap: 0K/1028M used/tot Maby it's just too much for Atom ? best regards M.K. W dniu 2010-11-05 20:20, Michay Koc pisze: Thank You for your time. The patch seems to resolve both problems on Atom platform. Will check Core2Duo later. Thanks once again Best regard M.K. W dniu 2010-11-05 18:36, Bob Beck pisze: Are you able to try the following? see if it solves your problem. Index: sys/kern/vfs_bio.c === RCS file: /cvs/src/sys/kern/vfs_bio.c,v retrieving revision 1.126 diff -u -r1.126 vfs_bio.c --- sys/kern/vfs_bio.c 3 Aug 2010 06:30:19 - 1.126 +++ sys/kern/vfs_bio.c 5 Nov 2010 17:32:44 - @@ -672,21 +672,10 @@ */ if (!ISSET(bp-b_flags, B_DELWRI)) { SET(bp-b_flags, B_DELWRI); - bp-b_synctime = time_uptime + 35; s = splbio(); reassignbuf(bp); splx(s); curproc-p_stats-p_ru.ru_oublock++;/* XXX */ - } else { - /* -* see if this buffer has slacked through the syncer -* and enforce an async write upon it. -*/ - if (bp-b_synctime time_uptime) { - bawrite(bp); - return; - } - } /* If this is a tape block, write the block now. */ if (major(bp-b_dev) nblkdev @@ -727,7 +716,6 @@ if (ISSET(bp-b_flags, B_DELWRI) == 0) { SET(bp-b_flags, B_DELWRI); - bp-b_synctime = time_uptime + 35; reassignbuf(bp); } } On 3 November 2010 05:17, Michay Kocm...@prime.pl wrote: Hi All, I've just upgraded two of my OpenBSD machines to 4.8: hw.machine=i386 hw.model=Intel(R) Core(TM)2 Duo CPU E8400 @ 3.00GHz (GenuineIntel 686-class) hw.product=DG31PR and hw.machine=i386 hw.model=Intel(R) Atom(TM) CPU D510 @ 1.66GHz (GenuineIntel 686-class) hw.product=D510MO Dmesgs are below. The problem is that they freeze every time I try to: - rsync two local filesystems on different physical disks - high disk IO - about 30GB - run nagios with about 900 probes - hight network IO and ndcpy like 3000 in systat, lots of forks, load average raising to 5 and above High disk IO freeze occurs about 30 seconds after rsync start and is permanent. High network IO freeze occurs several minutes after nagios start and sometimes machines are responsive for limited time. Pkill nagios resolves the problem, machine becomes responsive. In both cases machines behind nat still have internet connectivity. Local services like ssh or console are unavailable. Snapshot from 2010-11-02 22:51:00 does not resolve the issue. The Atom machine freezes much faster than Core2Duo. any help appreciated best regards M.K. Core2Duo dmesg: OpenBSD 4.8 (GENERIC.MP) #359: Mon Aug 16 09:16:26 MDT 2010 dera...@i386.openbsd.org:/usr/src/sys/arch/i386/compile/GENERIC.MP cpu0: Intel(R) Core(TM)2 Duo CPU E8400 @ 3.00GHz (GenuineIntel 686-class) 3.01 GHz cpu0: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,VMX,SMX,EST,TM2,SSSE3,CX16,xTPR,PDCM,SSE4.1 real mem = 3476889600 (3315MB) avail mem = 3410038784 (3252MB) mainbus0 at root bios0 at mainbus0: AT/286+ BIOS, date 02/27/08, SMBIOS rev. 2.4 @ 0xe8170 (42 entries) bios0: vendor Intel Corp. version PRG3110H.86A.0047.2008.0227.1745 date 02/27/2008 bios0: Intel Corporation DG31PR acpi0 at bios0: rev 2 acpi0: sleep states S0 S1 S3 S4 S5 acpi0: tables DSDT FACP APIC HPET MCFG acpi0: wakeup devices P0P1(S3) PS2K(S3) PS2M(S3) UAR1(S3) P0P2(S4) USB0(S3) USB1(S3) USB2(S3) USB3(S3) EUSB(S3) MC97(S4) PEX0(S4) PEX1(S4) PEX2(S4) PEX3(S4) SLPB(S4) PWRB(S3) acpitimer0 at acpi0: 3579545 Hz, 24 bits acpimadt0 at acpi0 addr 0xfee0: PC-AT compat cpu0 at mainbus0: apid 0 (boot processor) cpu0: apic clock running at 333MHz cpu1 at mainbus0: apid 1 (application processor) cpu1: Intel(R) Core(TM)2 Duo CPU E8400 @ 3.00GHz (GenuineIntel 686-class) 3 GHz cpu1: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,DS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,DS-CPL,VMX,SMX,EST,TM2,SSSE3,CX16,xTPR,PDCM,SSE4.1 ioapic0 at
xenocara: make release fails on vax/4.8-stable
Hi, On vax/4.8-stable, make release in /usr/xenocara fails with + install -c -o root -g wheel -m 644 /usr/xenocara/etc/X11.vax/xorg.conf /usr/dest/etc/X11 install: /usr/xenocara/etc/X11.vax/xorg.conf: No such file or directory *** Error code 71 Stop in /usr/xenocara (line 97 of Makefile). This seems to be fixed in revision 1.38 of /usr/xenocara/Makefile (commit comment: make 'make release' work on vax), but this is only in -current, not in 4.8-stable. Shouldn't this be commited to the stable branch as well? Thanks. Maurice
Re: relayd port to linux
On Fre 05.11.2010 10:45, Theo de Raadt wrote: due to the fact that openssh and some other parts of openbsd are ported to linux maybe you can tell me if you plan to make a openrelayd which is able to compile on linux. relayd depends deeply on pf. so the answer is no. ok, sorry for rush. Do you know a good replacement for stunnel with http-header rewrite on non openbsd OS?!
Re: relayd port to linux
On Fri, 5 Nov 2010 22:31:42 +0100, Aleksandar Lazic wrote: On Fre 05.11.2010 10:45, Theo de Raadt wrote: due to the fact that openssh and some other parts of openbsd are ported to linux maybe you can tell me if you plan to make a openrelayd which is able to compile on linux. relayd depends deeply on pf. so the answer is no. ok, sorry for rush. Do you know a good replacement for stunnel with http-header rewrite on non openbsd OS?! 1: Would you ask a linux mailing list for advice about a program to run on a non-linux OS? 2: Is your Google key broken? *** NOTE *** Please DO NOT CC me. I am subscribed to the list. Mail to the sender address that does not originate at the list server is tarpitted. The reply-to: address is provided for those who feel compelled to reply off list. Thankyou. Rod/ --- This life is not the real thing. It is not even in Beta. If it was, then OpenBSD would already have a man page for it.
Re: relayd port to linux
http://en.wikipedia.org/wiki/Reverse_proxy On Fri, Nov 5, 2010 at 5:31 PM, Aleksandar Lazic al-open...@none.at wrote: On Fre 05.11.2010 10:45, Theo de Raadt wrote: due to the fact that openssh and some other parts of openbsd are ported to linux maybe you can tell me if you plan to make a openrelayd which is able to compile on linux. relayd depends deeply on pf. so the answer is no. ok, sorry for rush. Do you know a good replacement for stunnel with http-header rewrite on non openbsd OS?!
Font Rendering Issue on 4.8-release
Hey Misc, I just did a fresh install of 4.8 to celebrate the release but I'm having some issues with fonts rendering in X (using Openbox as my wm). I've seen the issues in a few different applications, including Firefox, gnome-terminal, xterm and claws-mail. I don't know all fonts that are having issues, but currently: * the 'g' in Courier * the 'A' in Monospace * the 'u' and 'G' in serif The characters rendering incorrectly do not seem to be consistent between restarts of X. Any advice would be greatly appreciated. Thanks! -Sky
Re: relayd port to linux
On Fri, Nov 05, 2010 at 10:31:42PM +0100, Aleksandar Lazic wrote: On Fre 05.11.2010 10:45, Theo de Raadt wrote: due to the fact that openssh and some other parts of openbsd are ported to linux maybe you can tell me if you plan to make a openrelayd which is able to compile on linux. relayd depends deeply on pf. so the answer is no. ok, sorry for rush. Do you know a good replacement for stunnel with http-header rewrite on non openbsd OS?! You could run openbsd and be done with it. Unlike linux is doesn't suck so that helps that decision.
Re: diskmap(4) interface and live USB fstab file
fwiw, in -current, USB attach order should be quite predictable. there are no longer multiple threads attaching USB devices. attachment is now done in a single thread, and it is done in the same order every time. of course, if you change which USB ports the devices are connected to between boots, or disconnect/reconnect while booted then the order might change. On Fri, Nov 05, 2010 at 08:37:37PM +0800, Marcus wrote: http://www.openbsd.org/faq/faq1.html#WhatsNew says: diskmap(4) interface People using USB attached storage or softraid(4) configurations often had difficulty with drive identifiers changing from boot to boot, or between hardware configurations. diskmap(4) allows you to mount drives by unique disklabel UIDs rather than how they are attached, so now you can use the same /etc/fstab on your USB flash disk without worrying wheter it would come up as sd0, sd1 sd2, etc. http://www.openbsd.org/faq/faq14.html#flashmemLive says: If your target machine has an ahci(4) or SCSI interface, you will probably find your USB drive's identifier changing. Having multiple versions of your /etc/fstab file may make this easier to fix (in single user mode). ---Question Would somebody rewrite #flashmemLive section for the diskmap interface change? or how to edit the /etc/fstab for live USB device without worrying wheter it would come up as sd0, sd1 sd2, etc. -- jake...@sdf.lonestar.org SDF Public Access UNIX System - http://sdf.lonestar.org
Seminario Nacional Compradores 2010, Puerto Vallarta 22 y 23 de Noviembre
[IMAGE] PRESENTA De la Planeacisn al Control Seminario Nacional Compradores 2010 Puerto Vallarta 22 y 23 de Noviembre PMS Capacitacisn Efectiva de Mixico con el compromiso de presentar a usted estrategias y herramientas que proporcionen a su empresa resultados inmediatos, presentamos este exclusivo Seminario-Taller diseqado para que todo responsable del area de compras pueda dominar y convertir el departamento en un centro de ganancias que contribuya a los resultados del negocio. El Lic. Ariel Valero especialista de PMS de Mixico quiin por su trayectoria y experiencia en el ramo guiara a cada participante a desarrollar y aterrizar cada uno de los puntos que se veran en estos dos dmas de trabajo. OBJETIVOS: Identificar y destacar los tipos y principios aplicados en los distintos niveles de planeacisn Conocer las opciones y diagnosticar el enfoque estratigico de la actual estructura organizacional del departamento Definir los criterios de interaccisn interna y externa que optimicen la influencia estratigica en su funcisn Destacar los elementos objetivos que faciliten la medicisn de resultados DIRIGIDO A: Gerentes y Supervisores de Abastecimientos, Compradores, Lmderes de Proyecto, Responsables de Materiales, Coordinadores de la cadena de Abastecimientos y personal involucrado en estructuracisn de las funciones de Adquisiciones. Mayores informes responda este correo electrsnico con los siguientes datos. Empresa: Nombre: Telifono: Email: Nzmero de Interesados: Y en breve le haremos llegar la informacisn completa del evento. O bien comunmquense a nuestros telifonos un ejecutivo con gusto le atendera Tels. (33) 8851-2365, (33)8851-2741. Copyright (C) 2010, PMS Capacitacisn Efectiva de Mixico S.C. Derechos Reservados. PMS de Mixico, El logo de PMS de Mixico son marcas registradas. ADVERTENCIA PMS de Mixico no cuenta con alianzas estratigicas de ningzn tipo dentro de la Republica Mexicana. NO SE DEJE ENGAQAR - DIGA NO A LA PIRATERIA. Todos los logotipos, marcas comerciales e imagenes son propiedad de sus respectivas corporaciones y se utilizan con fines informativos solamente. Este Mensaje ha sido enviado a misc@openbsd.org como usuario de Pms de Mixico o bien un usuario le refiris para recibir este boletmn. Como usuario de Pms de Mixico, Si usted ha recibido este mensaje por error, haga caso omiso de el y reporte su cuenta respondiendo este correo con el subject BAJAcompras Unsubscribe to this mailing list, reply a blank message with the subject UNSUBSCRIBE BAJAcompras Tenga en cuenta que la gestisn de nuestras bases de datos es de suma importancia y no es intencisn de la empresa la inconformidad del receptor. [demime 1.01d removed an attachment of type image/jpeg which had a name of image003.jpg]
Re: SSH Connection Accounting
Thanks for the reply, but what I hope to find was an administrative tool rather than resource/building blocks to build such an application. On Thu, Nov 4, 2010 at 2:46 PM, Jan Stary h...@stare.cz wrote: On Nov 04 14:35:12, Tito Mari Francis Esca??o wrote: Good day! I'm googling for resources on SSH connection accounting and unfortunately for me, I keep bumping into Cisco-related resources online. What I need to know is how to determine the SSH connection accounting per user, like if a user is currently connected, when a user connected, how long the SSH session took if successfully connected and maybe other related information. I hope you can provide me pointers on this. Thanks! man man man -k accounting
OpenBSD bridge setup
Problem Description: I'm trying to filter VLANs on the bridge. However, when enabling VLAN devices on the em1 interface the bridge does not work. Test Setup: The 2910AL-24G port 19 has its ports configured as TAGGED for VLAN 300 and VLAN 302 with no other VLANs are enabled on this port. This cable enters the bridge via em0 of the bridge and em1 connects to port 1 on the HP5304XL which is configured for TAGGED VLAN 300 and VLAN 302. Port two is configured as VLAN 300 UNTAGGED. HP2910AL-24G (port 19) --- OpenBSD Bridge --- HP 5304XL (port 1) OS - OpenBSD 4.8-beta (GENERIC.MP) #259: Tue Aug 3 09:06:37 MDT 2010 (no difference with newer versions) PF - Disabled Two physical interfaces em0 em1 VLAN devices # cat /etc/hostname.vlan300 vlan 300 vlandev em1 # cat /etc/hostname.vlan302 vlan 302 vlandev em1 cat /etc/hostname.em0 up cat /etc/hostname.em1 up Working configuration but without filtering. = cat /etc/hostname.bridge0 add em0 add em1 up With this configuration and no VLAN devices created the bridge works and the tags are passed appropriately, however I am unable to filter the traffic on the VLANs. dhclient eth0 on client works fine pinging out works fine Non-Working configuration with hopes of filtering == However, as soon as I create the vlan300 devices with a parent of em1 the bridge stops functioning and the client on HP5304XL Port 2 (UNTAGGED VLAN 300) stops functioning. This remains the same even if I add the vlan300 and vlan302 devices to the bridge. dhclient stops working ping is dead I'm stumped here. Any ideas? -- James A. Peltier Systems Analyst (FASNet), VIVARIUM Technical Director Simon Fraser University - Burnaby Campus Phone : 778-782-6573 Fax : 778-782-3045 E-Mail : jpelt...@sfu.ca Website : http://www.fas.sfu.ca | http://vivarium.cs.sfu.ca http://blogs.sfu.ca/people/jpeltier MSN : subatomic_s...@hotmail.com
Re: diskmap(4) interface and live USB fstab file
On 11/05/10 08:37, Marcus wrote: ---Question Would somebody rewrite #flashmemLive section for the diskmap interface change? or how to edit the /etc/fstab for live USB device without worrying wheter it would come up as sd0, sd1 sd2, etc. ---Answer: Somebody will, as time permits. You do realize, though: you could be...somebody. However, this needs to be a whole new section, as the diskmap stuff is useful for a whole lot more than just flash disks. Then it needs to be linked in to a few other places in the FAQ. It requires a fair amount of explanation about what happens and why, not blind copy/paste stuff, and present both the new disk and existing disk cases. This IS on my very near-term list, as it is the current hold-up for a softraid FAQ entry that keeps getting held up as things evolve (for the better!). It's a very cool feature, and changes a lot of things. But orders without good diffs attached (er..inlined. :) don't help. Nick.
Re: relayd port to linux
On 11/05/2010 05:31 PM, Aleksandar Lazic wrote: On Fre 05.11.2010 10:45, Theo de Raadt wrote: due to the fact that openssh and some other parts of openbsd are ported to linux maybe you can tell me if you plan to make a openrelayd which is able to compile on linux. relayd depends deeply on pf. so the answer is no. ok, sorry for rush. Do you know a good replacement for stunnel with http-header rewrite on non openbsd OS?! Well, besides Marco being right about the best Unix system for networking out there (OpenBSD, keep in mind I manage a lot of reenucksh systems too), I would check out nginx or mod_proxy_balancer. I am big into puppet (uses ssl for communication), and I load balance with mod_proxy_balancer, and I know a lot of people who use nginx (but not me). -- -- Joe McDonagh Operations Engineer AIM: YoosingYoonickz IRC: joe-mac on freenode When the going gets weird, the weird turn pro.