Re: [osol-discuss] zone IPS issues after Solaris 11 Express - Solaris 11 release upgrade
On Mon, Nov 28, 2011 at 1:12 PM, Brett Eisenberg br...@librum.org wrote: It would appear that there are some strange IPS issues within after upgrading from 11E to 11R. If anyone has any pointers on possible paths forward, it would be much appreciated. Best, Brett root@global:~# zoneadm -z zone1 halt root@global:~# zoneadm -z zone1 detach root@global:~# zoneadm -z zone1 attach -u This would have brought the OS bits in sync between the global zone and zone1. Progress being logged to /var/log/zones/zoneadm.2028T180910Z.zone1.attach Attaching... Installing: Using existing zone boot environment Zone BE root dataset: data/zones/zone1/rpool/ROOT/zbe-8 Cache: Using /var/pkg/publisher. Updating image format Image format already current. Updating non-global zone: Linking to image /. Updating non-global zone: Syncing packages. No updates necessary for this image. Updating non-global zone: Zone updated. Result: Attach Succeeded. Log saved in non-global zone as /zones/zone1/root/var/log/zones/zoneadm.2028T180910Z.zone1.attach root@global:~# zoneadm -z zone1 boot root@global:~# zlogin zone1 [Connected to zone 'zone1' pts/12] Oracle Corporation SunOS 5.11 11.0 November 2011 root@zone1:~# pkg list -u NAME (PUBLISHER) VERSION IFO consolidation/SunVTS/SunVTS-incorporation 0.5.11-0.172.0.0.0.0.0 i-- consolidation/ips/ips-incorporation 0.5.11-0.175.0.0.0.2.2576 i-- consolidation/osnet/osnet-incorporation 0.5.11-0.175.0.0.0.2.1 i-- entire 0.5.11-0.175.0.0.0.2.0 i-- package/pkg 0.5.11-0.175.0.0.0.2.2576 i-- system/core-os 0.5.11-0.175.0.0.0.2.1 i-- system/file-system/nfs 0.5.11-0.175.0.0.0.2.1 i-- system/file-system/zfs 0.5.11-0.175.0.0.0.2.1 i-- system/kernel 0.5.11-0.175.0.0.0.2.1 i-- system/zones 0.5.11-0.175.0.0.0.2.1 i-- root@zone1:~# pkg update Creating Plan \ pkg update: No solution was found to satisfy constraints Plan Creation: Package solver has not found a solution to update to latest available versions. This may indicate an overly constrained set of packages are installed. latest incorporations: pkg://solaris/consolidation/SunVTS/SunVTS-incorporation@0.5.11,5.11-0.175.0.1.0.4.0:2008T154138Z Dependency analysis is unable to determine exact cause. Try specifying expected results to obtain more detailed error messages. Solaris 11 introduces parent dependencies, as described on page 38 of the draft Packaging and Delivering Software with the Image Packaging System available at http://defect.opensolaris.org/bz/attachment.cgi?id=4607. I believe that what pkg is inelegantly telling you is that you can't update to the latest version of the desired packages because that would break the parent dependencies. That is, unlike Solaris 11 Express, Solaris 11 has mechanisms in place to be sure that the global zone and non-global zone software that must be in sync is in sync. If you want the zone to run the latest version of the packages that make up Solaris, the global zone must also be updated. If you need a more detailed or certain answer, you may want to ask on pkg-disc...@opensolaris.org. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Solaris11 zone networking on VirtualBox
On Tue, Nov 15, 2011 at 12:14 AM, Deniz Rende deniz.re...@gmail.com wrote: Hi Mike, Option 1: Use shared-stack But that means that I am not going to be able to use all the goodies that come with ip-excl. :( Agreed. :( You will then need to transition your network configuration in the zone from net0 to net1. Do you mean, sysconfig configure here or some othe procedure. Right now it looks like this: LINK CLASS MTU STATE OVER net0 phys 1500 up -- net1 phys 1500 unknown -- zdev1/net1 phys 1500 unknown -- Whatever you do to reconfigure it, you need to do it in the zone. You can use sysconfig configure, or use ipadm. I think the steps are something along the lines of: root@zone# ipadm delete-ip net0/something root@zone# ipadm create-ip net1/v4 root@zone# ipadm create-addr -T static -a local=addr/mask net1/v4 My fingers are still learning the new magic, so there could be something horribly wrong in my advice. If Seb gives conflicting advice, he's probably right. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Change host name on ora-Sol 11
On Tue, Nov 15, 2011 at 8:43 PM, Harry Putnam rea...@newsguy.com wrote: Is there some different way to change host name in oracle-solaris 11? There is no /etc/nodename file. and I find the only place under /etc/ the hosts' current name is referenced is in: /etc/svc/profile/site/sc_profile.xml That doesn't look like the kind of place renaming would be done. So any special technique involved or am expected to create /etc/nodename? From nodename(4): When a machine is standalone or its IP address is configured locally, the svc:/system/identity:node SMF service has a property, namely nodename, for setting the system name. To set the name of the node use the following command: # svccfg -s node setprop config/nodename = mynodename # svcadm refresh node -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Solaris11 zone networking on VirtualBox
On Mon, Nov 14, 2011 at 10:46 PM, Deniz Rende deniz.re...@gmail.com wrote: root@solaris11:~# zonecfg -z zdev1 info zonename: zdev1 zonepath: /zones/zdev1 brand: solaris autoboot: true bootargs: -m verbose file-mac-profile: pool: limitpriv: scheduling-class: ip-type: exclusive hostid: fs-allowed: anet: linkname: net0 lower-link: auto It could be that this anet resource automatically picked the wrong lower link. From the global zone, what do dladm show-link and ipadm show-addr say? -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Solaris11 zone networking on VirtualBox
On Mon, Nov 14, 2011 at 11:25 PM, Deniz Rende deniz.re...@gmail.com wrote: It could be that this anet resource automatically picked the wrong lower link. From the global zone, what do dladm show-link and ipadm show-addr say? I get the following information: root@solaris11:~# dladm show-link LINK CLASS MTU STATE OVER net0 phys 1500 up -- zdev1/net0 vnic 1500 up net0 root@solaris11:~# ipadm show-addr ADDROBJ TYPE STATE ADDR lo0/v4 static ok 127.0.0.1/8 net0/v4 static ok 192.168.1.20/24 lo0/v6 static ok ::1/128 net0/v6 addrconf ok fe80::a00:27ff:fe22:cf2f/10 I don't know if it worths to mention but the version of Solaris 11 is the text installI also don't have any trouble GZ networking: root@solaris11:~# ping www.google.com www.google.com is alive I suspect that what is happening is that the zdev1/net0 vnic thinks it is OK to allocate another MAC (Ethernet) address but virtualbox isn't happy with it and thinks something in the VM is trying to spoof packets There may be a way within virtualbox to disable this behavior, but I've not used virtualbox recently enough to remember. Assuming you can't disable the vbox feature that seems to be getting in the way, instead of using an anet resource which creates a vnic on top of the global zone's net0, try one of the following: Option 1: Use shared-stack zonecfg -z zdev1 remove anet linkname=net0 set ip-type=shared add net set name=net0 set addr=192.168.1.13/24 set defrouter=192.168.1.254 end exit zoneadm -z zdev1 reboot You will probably need to do some ipadm magic (ipadm delete-ip net0/something?) to make the configuration for net0 go away, else network/physical:default will probably be in maintenance. Option 2: Allocate another network In this option, you need to have virtualbox provide another NIC. That is, shut down the VM, within the VirtualBox GUI, configure a second network to look just like the first one, then boot the VM. Once you are back in Solaris: zonecfg -z zdev1 remove anet linkname=net0 add net set physical=net1 end exit zoneadm -z zdev1 reboot You will then need to transition your network configuration in the zone from net0 to net1. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] snv_134 rebooting without any cause
On Fri, Apr 29, 2011 at 7:11 AM, Orkan Dere orkand...@gmail.com wrote: Hello, I have a Sun Fire X4540 running opensolaris snv_134 and acting as nfs and iSCSI storage. It's rebooting irregularly (couldn't find any pattern). /var/adm/messages doesn't show any errors other than these (which I don't think is the cause): Apr 29 14:27:22 thor console-kit-daemon[482]: [ID 702911 daemon.warning] GLib-GObject-WARNING: g_object_set_property: construct property seat-id for object `CkSession' can't be set after construction Apr 29 14:27:23 thor gnome-session[1668]: [ID 702911 daemon.warning] WARNING: IceListenForConnections returned 2 non-local listeners: inet/thor:54283,inet6/thor:47443 svcs -xv doesn't return anything so everything seems to be working fine. And I cannot figure out what's causing the reboots. Can you please help me to figure this out? I suspect you have a hardware, cooling, or power problem. I suspect you may find more by poking around in the ILOM. http://download.oracle.com/docs/cd/E19469-01/820-5549-12/ -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Dropped my PC, S11e behaves strange??? Help!
On Mon, Mar 14, 2011 at 2:26 PM, Orvar Korvar knatte_fnatte_tja...@yahoo.com wrote: SunOS Release 5.11 Version snv_151a 64-bit Copyright (c) 1983, 2010, Oracle and/or its affiliates. All rights reserved. Hostname: JamesBond mount: No such file or directory svc:/system/filesystem/local:default: WARNING: /sbin/mountall -l failed: exit status 32 Mar 14 18:23:07 svc.startd[9]: svc:/system/filesystem/local:default: Method /lib/svc/method/fs-local failed with exit status 95. Mar 14 18:23:07 svc.startd[9]: system/filesystem/local:default failed fatally: transitioned to maintenance(see 'svcs -xv' for details) Any time you see a message about svc.startd reporting some error, the way to find out more about the error is: svcs -x service In this instance: svcs -x svc:/system/filesystem/local:default It will include a line that says where a log file is that will give you more information. Read the last few lines of that log file to see if it is helpful. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Dropped my PC, S11e behaves strange??? Help!
On Mon, Mar 14, 2011 at 4:29 PM, Orvar Korvar knatte_fnatte_tja...@yahoo.com wrote: Ok, I think it seems easier to zpool import than open the chassi again. I will try to boot live CD S11e and do a zpool import. I'll report back in 15 min. If it made it far enough to tell you that filesystem/local is in maintenance and gave you a login prompt, you can probably just log in, look at the log I previously suggested and fix the problem. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Solaris 11 : The new pkg command does not take http_proxy shell variable
On Wed, Feb 23, 2011 at 9:09 PM, Nitin Verma nitinshe...@gmail.com wrote: The new pkg command does not take http_proxy shell variable, as mentioned in the Man page. I am running Solaris 11 as guest in VBox. I have set http_proxy=http://my_proxy_ip:8080; when I run, # pkg refresh or # pkg set-publisher it just times out network unreachable!!! Any help would be appreciated! Please note: using curl with same proxy I can access internet from my Sol-11 VM. If the publisher(s) you are trying to access are https (instead of http) you will also need https_proxy set. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] can you help about process?
On Thu, Jan 13, 2011 at 2:27 PM, Deric Kwok deric.kwok2...@gmail.com wrote: Hi I can't find the top in the solairs. but how can I know the process using? I only know the system is using high as w load average: 55.55, 53.77, 49.33 prstat is what you are looking for. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Working with Cron under Solaris 10
On Fri, Jan 14, 2011 at 4:32 PM, Forrest Aldrich for...@gmail.com wrote: Running crontab -e is not applicable to what I'm doing here. I experimented with it and it turns out you can't use that functionality in a scripted environment, it doesn't work. I have a script that edits a line of a crontab, either commented out or not. Then it needs to tell cron to re-read the tabs. How about: tmpfile=`mktemp` crontab -l $tmpfile ... # munge the file here or in a pipeline with the previous command crontab $tmpfile rm -f $tmpfile Before I knew about that approach, I also did things like setting EDITOR (be sure VISUAL is unset) to the name of a script that knows how to munge it. That is, something like the following: env -i PATH=$PATH EDITOR=/my/script crontab -e And if you already edited it in place (I advise against it...) env -i PATH=$PATH EDITOR=touch crontab -e -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] How to compile libc or pthread ?
On Mon, Nov 29, 2010 at 3:26 PM, Afantee Lee dongli...@gmail.com wrote:
Thanks for the suggestion.
it seems Dtrace can profile the enter point and exit point of the program.
It is a nice tool, but, I probably need more insides into the pthread
function it self.
for example, I need to profile how many times the pthread_lock try to get the
access to the critical section. it is part of the pthread detail
implementation code.
please let me know if you know any other tool might be helpful.
Thank you a lot :)
Using the dtrace PID provider you can trace every instruction.
Consider the following:
Use mdb to disassemble a function:
$ echo pthread_rwlock_wrlock::dis | mdb /lib/libc.so.1
pthread_rwlock_wrlock: pushl %ebp
pthread_rwlock_wrlock+1:movl %esp,%ebp
pthread_rwlock_wrlock+3:pushl %ebx
pthread_rwlock_wrlock+4:subl $0x4,%esp__strdupa_len
pthread_rwlock_wrlock+7:call +0x0 pthread_rwlock_wrlock+0xc
pthread_rwlock_wrlock+0xc: popl %ebx
pthread_rwlock_wrlock+0xd: addl $0x9435c,%ebx_vwscanf_c89+0x14
pthread_rwlock_wrlock+0x13: subl $0x8,%esp
pthread_rwlock_wrlock+0x16: pushl $0x0
pthread_rwlock_wrlock+0x18: pushl 0x8(%ebp)
pthread_rwlock_wrlock+0x1b: call -0x1d0 rw_wrlock_impl
pthread_rwlock_wrlock+0x20: addl $0x14,%esp
pthread_rwlock_wrlock+0x23: popl %ebx
pthread_rwlock_wrlock+0x24: leave
pthread_rwlock_wrlock+0x25: ret
Use dtrace to list the available probes in that function:
# dtrace -l -n 'pid$target:libc.so.1:pthread_rwlock_wrlock:' -c /bin/true
ID PROVIDERMODULE FUNCTION NAME
71738pid6477 libc.so.1 pthread_rwlock_wrlock return
71739pid6477 libc.so.1 pthread_rwlock_wrlock entry
71740pid6477 libc.so.1 pthread_rwlock_wrlock 0
71741pid6477 libc.so.1 pthread_rwlock_wrlock 1
71742pid6477 libc.so.1 pthread_rwlock_wrlock 3
71743pid6477 libc.so.1 pthread_rwlock_wrlock 4
71744pid6477 libc.so.1 pthread_rwlock_wrlock 7
71745pid6477 libc.so.1 pthread_rwlock_wrlock c
71746pid6477 libc.so.1 pthread_rwlock_wrlock d
71747pid6477 libc.so.1 pthread_rwlock_wrlock 13
71748pid6477 libc.so.1 pthread_rwlock_wrlock 16
71749pid6477 libc.so.1 pthread_rwlock_wrlock 18
71750pid6477 libc.so.1 pthread_rwlock_wrlock 1b
71751pid6477 libc.so.1 pthread_rwlock_wrlock 20
71752pid6477 libc.so.1 pthread_rwlock_wrlock 23
71753pid6477 libc.so.1 pthread_rwlock_wrlock 24
71754pid6477 libc.so.1 pthread_rwlock_wrlock 25
Use dtrace to see how frequently each of those instructions are hit:
# dtrace -qn 'pid$target:libc.so.1:pthread_rwlock_wrlock: {
@freq[probefunc,probename] = count() } tick-1s {printf(%Y\n,
walltimestamp); printa(%...@d %s:%s\n, @freq); trunc(@freq)}' -p 3395
2010 Nov 30 15:36:36
2010 Nov 30 15:36:37
2010 Nov 30 15:36:38
2010 Nov 30 15:36:39
12 pthread_rwlock_wrlock:0
12 pthread_rwlock_wrlock:1
12 pthread_rwlock_wrlock:13
12 pthread_rwlock_wrlock:16
12 pthread_rwlock_wrlock:18
12 pthread_rwlock_wrlock:1b
12 pthread_rwlock_wrlock:20
12 pthread_rwlock_wrlock:23
12 pthread_rwlock_wrlock:24
12 pthread_rwlock_wrlock:25
12 pthread_rwlock_wrlock:3
12 pthread_rwlock_wrlock:4
12 pthread_rwlock_wrlock:7
12 pthread_rwlock_wrlock:c
12 pthread_rwlock_wrlock:d
12 pthread_rwlock_wrlock:entry
12 pthread_rwlock_wrlock:return
2010 Nov 30 15:36:40
2010 Nov 30 15:36:41
2010 Nov 30 15:36:42
2010 Nov 30 15:36:43
2010 Nov 30 15:36:44
...
--
Mike Gerdts
http://mgerdts.blogspot.com/
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Solaris 11 Express is out
On Mon, Nov 15, 2010 at 1:03 PM, Karel Gardas karel.gar...@centrum.cz wrote: interesting would be to know if the upgrade from OS 2009.06 is working well or not. I haven't tried that exact path, but I have image-update'd from OpenIndiana to something that looks quite similar to Solaris 11 Express (build 150 instead of 151a) I've been using the machine on which I did this as my primary machine for a couple weeks now without any problems. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] System always hits dns
On Sat, Nov 6, 2010 at 4:19 AM, Milan Jurik milan.ju...@oracle.com wrote: Hi Mike, Mike Gerdts píše v pá 05. 11. 2010 v 13:06 -0500: On Fri, Nov 5, 2010 at 12:22 PM, Milan Jurik milan.ju...@oracle.com wrote: Hi, Mike Gerdts píše v pá 05. 11. 2010 v 12:10 -0500: It is trying to do an IPv6 lookup (via ipnodes) before trying an IPv4 lookup (via hosts). Since you probably aren't using IPv6, you can get the desired behavior with: hosts: files nis dns ipnodes: files such config is really bd idea. Keep them in sync from Solaris 10 releases. Can you elaborate? E.g. double parsing of flat file before going to DNS? Before PSARC/2006/299 most people that would find this configuration desirable would have a very small ipnodes file. That is, it would likely only have the entries added by the installer. If IPv6 is not in use, having ipnodes go to DNS does little other than cause unnecessary DNS requests that will always fail. My observation is that the times that you notice it is failing are those times where it causes a really long timeout to be invoked. In correctly set environment DNS server will not timeout for records requests. Not everyone is so lucky to have a properly working DNS server. In particular, if you are having network problems that prevent you from reaching the DNS server(s) then it can be quite frustrating. There are cases where such config will bring something good but only in rare cases it is not hiding some real setup problem. There is still not clear why with common setup of hosts/ipnodes he does not see effect of nscd cache and queries going to DNS even if needed data are in files. Perhaps cached negative lookups (misses for IPv6) are timing out before positive lookups? That is, it forgets that ipnodes doesn't provide the answer before it forgets that hosts has the answer. As a result, it has to go look to ipnodes since it is always searched before hosts. Still, ipnodes are not obsolete because if you query only hosts database, DNS backend will generate only query only fo A record, in case of ipnodes, it will generate both (so your workaround will work). My words were not chosen carefully... Turning /etc/inet/ipnodes into a symbolic link to /etc/inet/hosts for backward compatibility, as was done with the implementation PSARC/2006/299 makes ipnodes(4) obsolete in my mind. Solaris considers it a stable (not obsolete) interface. The effect of PSARC/2006/299 is that the non-existence of the ipnodes file a completely workable configuration aside from those users that would actually modify /etc/inet/ipnodes and expect it to have some effect. I did not intend to say that the existence of ipnodes in nsswitch.conf was obsolete. Sorry for the confusion. http://arc.opensolaris.org/caselog/PSARC/2006/299/materials/spec.txt http://hg.genunix.org/onnv-gate.hg/rev/318aeb1049d9 How does one force software that they did not write to use only hosts lookups? I've observed this problematic behavior on systems that do not have and have never had an IPv6 address configured. I'm not aware of applications that first look to see if it is possible to speak IPv6 before asking the name service switch for an IPv6 address. Initially I was thinking it would be a useful enhancement to make the name service switch fail fast if IPv6 was not configured and someone was doing an ipnodes lookup that would hit DNS. However, it is quite reasonable to expect that an IPv4-only host would be able to ask about the IPv6 world. What I would really like is to say: ipnodes: none For the case where IPv6 is not used. My first attempt at solving this problem was to remove/comment the ipnodes nsswitch.conf entry. However, this seemed to have the effect of ipnodes choosing the same name services as hosts. My initial debugging of these unwanted IPv6 lookups dates back many years (Solaris 8 or 9?) so things may have changed. Best regards, Milan Likewise, Mike -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] System always hits dns
It is trying to do an IPv6 lookup (via ipnodes) before trying an IPv4 lookup (via hosts). Since you probably aren't using IPv6, you can get the desired behavior with: hosts: files nis dns ipnodes: files Releases of Solaris/OpenSolaris later than what you appear to be running know how to look up IPv6 addresses via hosts as well. That is, ipnodes in nsswitch.conf becomes obsolete. Mike On Fri, Nov 5, 2010 at 12:03 PM, Mike DeMarco mikej...@yahoo.com wrote: Build 134 My nsswitch.conf file is setup as such hosts files nis dns ipnodes files nis dns when snooping the interface I see dns lookups even though the ipaddress lives in the local host file and or the nis database. It was a very early problem with nsswitch that it would still follow through with requests even though they were satisfied with the prior request. I tried to limit it going out to the next resource if the current one was found with hosts: files [SUCCESS=return] nis [SUCCESS=return] dns ipnodes:files [SUCCESS=return] nis [SUCCESS=return] dns But still every request goes to dns no matter what it finds in the files database or nis database. Waiting for a dns timeout on every ssh to a host that is not in dns is a killer. Anyone seen this behavior and have input. Thanks -- This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] System always hits dns
On Fri, Nov 5, 2010 at 12:22 PM, Milan Jurik milan.ju...@oracle.com wrote: Hi, Mike Gerdts píše v pá 05. 11. 2010 v 12:10 -0500: It is trying to do an IPv6 lookup (via ipnodes) before trying an IPv4 lookup (via hosts). Since you probably aren't using IPv6, you can get the desired behavior with: hosts: files nis dns ipnodes: files such config is really bd idea. Keep them in sync from Solaris 10 releases. Can you elaborate? If IPv6 is not in use, having ipnodes go to DNS does little other than cause unnecessary DNS requests that will always fail. My observation is that the times that you notice it is failing are those times where it causes a really long timeout to be invoked. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] [b 133] Catching messages on reboot from failure
On Wed, Oct 27, 2010 at 9:16 AM, Harry Putnam rea...@newsguy.com wrote: Mike Gerdts mger...@gmail.com writes: In which case you can edit the kernel line in grub to add a -k option to the boot options. If the system panics, you will be dropped to a kmdb prompt. You can manually enter kmdb with F1-A or shift-break from a text console. You can use Ctrl-Alt-F1 to shift to the text console if you aren't there already. Once at the kmdb prompt, you can use ::msgbuf to see the things that have scrolled off the screen. You should get the output a page at a time. If you need to provide this Thanks,, very helpful info there I'm a bit confused though is all of this dependent on having the system panic, or do you mean you can enter kmbd from a text console any time? For example, if I were to edit /rpool/boot/grub/menu.lst at this line: kernel$ /platform/i86pc/kernel/$ISADIR/unix -B $ZFS-BOOTFS # console=graphics by adding `-k' before the octothorpe above. On reboot, and from a text console I could enter kmbd as you describe? Yes Oh, and by the way, what is `shift-break'... is it on a standard us qwerty keyboard? Maybe the `pause' key? Quite likely. Your initial description implied to me that the system was in some state where you were stuck with just the text console, potentially locked up. The same text that is available with ::msgbuf in mdb is also available from the dmesg command. The dmesg command can be run from a more capable terminal than the console - such as from a GUI login session or remotely through ssh. Most of the output that goes to the console will also be logged by syslog - which I see you found in a message you sent just a couple minutes ago. If the system is behaving well and you want to poke around in it without pausing the kernel, you can use mdb -k as well. This can be run from any root (or otherwise appropriately privileged) shell - you don't have to be on the console. The use of mdb -k on a live system does not require that it was booted with the -k option. Also, if you can find services that were unable to start with svcs -xv. Each of them will have a log file (see output of svcs -xv) that may have more useful information. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] [b 133] Catching messages on reboot from failure
On Wed, Oct 27, 2010 at 9:31 AM, Harry Putnam rea...@newsguy.com wrote: Marion Hakanson hakan...@ohsu.edu writes: As someone else mentioned, most of this stuff will end up in the /var/adm/messages file. Except, of course, for the very interesting lowest-level boot problems. yeah... I got a little carried away there... I had long ago setup a /var/adm/debug.log that catches everything log message issued. with *.debug /var/adm/debug.log in /etc/syslog.conf Then forgot I had done so... The messages are there alright but now I'm finding they are not all that helpful... an example at the end. The reason why it's so primitive is that at this point (in Solaris, OpenSolaris, Linux, or Windows), you're still pretty much dependent on whatever features the BIOS provides you, which limits you just to the old-fashioned, low-tech text-only console. There's just not enough software running yet to give you a GUI. I'm not sure what you mean here With the same bios, linux manages to provide a text console that is vastly more useful... ie, it has gpm running. I wasn't asking for a gui, I like a text console... but do like to be able to mouse copy from the buffer. About the console messages; they were concerning how to find the output using fmdump, and using the `Event ID' to do so. Now armed with the Event IDs that had scrolled off the screen: I still learn nothing very useful, that is, unless by pounding away on man fmdump and probably a number of other man pages this can be made sense of. But not at all obvious what any of this might mean: fmdump -v -u 4021154c-2d57-c54e-ae82-ea27fc2d19fa TIME UUID SUNW-MSG-ID Oct 14 04:14:02.1292 4021154c-2d57-c54e-ae82-ea27fc2d19fa ZFS-8000-GH 100% fault.fs.zfs.vdev.checksum Problem in: zfs://pool=z3/vdev=e35825af18775fc2 Affects: zfs://pool=z3/vdev=e35825af18775fc2 FRU: - Location: - It looks to me like you got a checksum error in a zpool named z3. Oct 25 17:06:24.7775 4021154c-2d57-c54e-ae82-ea27fc2d19fa FMD-8000-4M Repaired 100% fault.fs.zfs.vdev.checksum Repair Attempted Problem in: zfs://pool=z3/vdev=e35825af18775fc2 Affects: zfs://pool=z3/vdev=e35825af18775fc2 FRU: - Location: - It was repaired Oct 25 17:06:24.8987 4021154c-2d57-c54e-ae82-ea27fc2d19fa FMD-8000-6U Resolved 100% fault.fs.zfs.vdev.checksum Repair Attempted Problem in: zfs://pool=z3/vdev=e35825af18775fc2 Affects: zfs://pool=z3/vdev=e35825af18775fc2 FRU: - Location: - And it has been resolved - you don't have to worry about it any more. But, I would run zpool status z3 to confirm that it says everything is healthy. It may also be advisable to run zpool scrub z3 to have zfs look for any more problems (hopefully all correctable) before they pile up to a point where they aren't correctable. If you keep having problems, it may point to hardware problems (disk, cable, disk controller, memory, etc.). -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] [b 133] Catching messages on reboot from failure
On Tue, Oct 26, 2010 at 9:09 PM, Marion Hakanson hakan...@ohsu.edu wrote: rea...@newsguy.com said: Why is the console interface left so primitive?. Seems it would at least have a usable mouse so one could have some chance of copy paste when there are what appear to be important messages written to console. Hi Harry, As someone else mentioned, most of this stuff will end up in the /var/adm/messages file. Except, of course, for the very interesting lowest-level boot problems. In which case you can edit the kernel line in grub to add a -k option to the boot options. If the system panics, you will be dropped to a kmdb prompt. You can manually enter kmdb with F1-A or shift-break from a text console. You can use Ctrl-Alt-F1 to shift to the text console if you aren't there already. Once at the kmdb prompt, you can use ::msgbuf to see the things that have scrolled off the screen. You should get the output a page at a time. If you need to provide this information to mailing lists, take a somewhat low resolution picture of it (e.g. a 1 megapixel picture from your mobile phone is most likely quite adequate). kmdb allows you to do many other interesting things, such as looking at running processes (::ptree, ::ps, ...) looking as to why a particular process is hung (::pgrep hungprocess | ::walk thread | ::findstack -v). If you are inclined to dig into this further, I would suggest perusing the mdb manual. Pretty much everything that works with mdb -k works with kmdb. The key exception that I've noticed is the lack of the ! operator to pipe dcmd output to a shell command. Considering that the OS is stopped while you are at a kmdb prompt, it's not surprising that ! doesn't work. The serial console advice below is also quite helpful if you have suitable hardware. Unfortunately, many systems these days lack a serial port. I doubt (without testing - I may be quite wrong) that a serial port hanging off of a USB port will be a very poor/fragile console. What you need to solve such scrolling off the screen problems is something even lower tech: a serial (RS232, COM-port) console. Not all desktop PC BIOS'es can redirect their BIOS text to a COM port, but Solaris Linux can be told to do so for their system console input output. And grub itself can be told to do this as well. Then you hook up your troublesome machine's serial (COM) port to a working machine's serial port, fire up a terminal emulator (Windows hyperterm will work; On Linux/Unix I would use conserver, but tip will do in a pinch), and watch the console messages that way. Of course, getting the serial ports wired correctly is a bit of an art (you may need a null modem cable, for example); And there are some boot-time flags you enter via grub to tell whatever kernel you're booting to temporarily use a tty console. Telling Google something like solaris boot serial console turns up quite a few references. Regards, Marion ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Quick question about the future
On Thu, Oct 14, 2010 at 7:14 AM, john kroll jek0...@sbcglobal.net wrote: What recent builds? The publicly released builds for community development ?? The only publicly released builds for community development have come from the community. The last one that Sun/Oracle published was b134. I suggest that you take a look at OpenIndiana[.org] if you are looking for a community driven distribution that continues on in the tradition of OpenSolaris. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] OpenSolaris cancelled, to be replaced with Solaris 11 Express
On Wed, Sep 8, 2010 at 7:10 AM, usafverteran us...@yahoo.com wrote: That is nonsense and just tabloid talk. IBM just recently announced AIX 7 for their POWER 7 server line. Linux cannot in any way use the full capabilities of POWER architecture the way it is exploited with AIX. If IBM was to eliminate AIX, they would have to also kill their POWER architecture, which isn't going to happen. Their POWER servers have a good margin and they make money from them. There are also very large companies that run AIX and have no intention to turn over their AIX-powered servers to Linux on cheap x86 hardware. Linux doesn't have anywhere near the features that AIX can boast about. RHEL clustering is a joke at best, and Xen virtualization clustering on RHEL is about as pathetic as one can get. There isn't any way I would entrust my company to RHEL for mission critical workloads. Absolutely not. There is a complete fabrication that Linux is cheaper than AIX or Solaris. Well, if I want to setup an infrastructure to patch, install, and maintain servers, it won't cost me any money to use NIM or JumpStart. However, to get those features with Red Hat Satellite server, you will pay hundreds per machine. So if you have 1000 servers you need to provision with RHEL, then you just spent $600,000 - $700,000. When they come out with KVM in RHEL 6, if you want to manage them it will cost you a few hundred more per virtual. EVERY time one turns around, Red Hat is charging for some feature that is free with AIX. The analog to Jumpstart is Kickstart. Kickstart is free. On x86 hardware, a typical network-based installation consists of PXE (DHCP + TFTP) + HTTP, FTP, and/or NFS. There are options for doing it purely from a CD/DVD and other non-network media. It is free and has been part of the OS for over a decade. This is starting to sound rather like Jumpstart, isn't it? From 2000: http://www.redhat.com/docs/manuals/linux/RHL-6-Manual/ref-guide/ From 2010: http://www.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6-Beta/html/Installation_Guide/ch-kickstart2.html The up2date and/or similar commands have been around for a long time to automatically patch (upgrade RPMs). Sun's free options have varied over time and from my experience have been more of a PITA than up2date. I still prefer Solaris. I'm just calling BS on this argument. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] OS 2009.06: ntpd dependency on zone startup
On Sun, Sep 5, 2010 at 8:10 AM, Karel Gardas karel.gar...@centrum.cz wrote:
Hello,
is it possible to configure kind of dependency between NTP service startup
and zone startup? The issue here is that I'm using virtual router created in
one zone on the workstation and during the workstation boot NTP service is
not able to locate NTP servers since the network connection is not working
due to not yet booted router zone.
Thanks,
Karel
My first instinct is to add a dependency on svc:/system/zones:default.
However, I think there is likely a race condition in that this
service will likely be online before the zones are actually fully up.
In particular, svc:/system/zones:default in the global zone may come
online prior to the appropriate routing related service in the
non-global zone.
Another approach would be to introduce a transient service that calls
a start script that does something like:
#! /bin/ksh
zone=...
svc=...# svc:/network/routing/route:default ?
function zonestate {
zoneadm -z $1 list -p | awk -F: '{print $3}'
}
while [[ $(zonestate $zone) != running ]] ; do
echo Waiting for $zone to start
sleep 1
done
zlogin $zlone svcadm enable -t $svc
This new service could then list svc:/network/ntp:default as
dependent on this new service. An example of using dependent can
be seen in /var/svc/manifest/system/metainit.xml to cause
svc:/system/metainit:default to (e.g.) come up in single user mode.
--
Mike Gerdts
http://mgerdts.blogspot.com/
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Build 134 NIS server
On Fri, Sep 3, 2010 at 11:07 AM, Mike DeMarco mikej...@yahoo.com wrote: Am I missing something or has NIS server been pulled from build 134? Perhaps you need to: pkg install service/network/nis -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SUNWipkg-brand
On Tue, Aug 31, 2010 at 11:36 PM, adidas totaap...@gmx.com wrote: please tell me how to know the pkg history. I am new to unix. Type pkg history at the prompt. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SUNWipkg-brand
On Sat, Aug 28, 2010 at 11:17 AM, adidas totaap...@gmx.com wrote: I got update from package manager now. I installed it. But, I am unable to find anything in this website about it. Details are given below. A url is given, but I lost it. Summary: None Size: 77.30 kB Category: None Installed Version: 0.5.11,5.11-0.111 Latest Version: 0.5.11,5.11-0.111 Packaging Date: Wed Aug 26 18:56:54 2009 FMRI: pkg:/sunwipkg-br...@0.5.11,5.11-0.111:20090826T185654Z Repository: opensolaris.org Based on the display at http://pkg.opensolaris.org/release/en/index.shtml, the repo for build 111 (2009/06 has not been updated since June 4, 2010. Note that the Packaging Date listed above was from just over a year ago. What does pkg history tell you? -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] OpenSolaris cancelled, to be replaced with Solaris 11 Express
On Sun, Aug 15, 2010 at 6:38 PM, Erik Trimble erik.trim...@oracle.com wrote: Oh, and a side remark - I think it would make a bit more of a pronounceable name if we used IlluminOS (note the extra in). Also, the word is cool and appropriate. http://www.urbandictionary.com/define.php?term=illuminos Maybe illuminos describes a person contributing to Illumos, just as a Joyeur is a person working for for Joyent. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] upcoming web event:strategy for Oracle's Sun Servers, Storage and solaris
On Thu, Aug 5, 2010 at 10:57 PM, Paul Gress pgr...@optonline.net wrote: On 08/ 5/10 11:31 PM, Edward Martinez wrote: I think this may be important to solaris customers:) excerpt: Tuesday, August 10, 2010 9:00 a.m. – 10:00 a.m. PT Live Webcast Join John Fowler, Executive Vice President, Systems, for a live update on the strategy and roadmap for Oracle’s Sun servers, storage and complete systems including Oracle Solaris. http://www.oracle.com/webapps/events/EventsDetail.jsp?p_eventId=119649src=7011191src=7011191Act=6evite=6 Some up news above. Some down news below: Oracle chops Sun HPC jobs http://www.theregister.co.uk/2010/08/05/sun_hpc_cuts_systems/ Which makes this even more confusing... http://blogs.sun.com/templedf/entry/not_dead_yet Does this mean that Grid Engine is shifting its focus from HPC to business computing? I've long thought that business computing could learn a thing or three from HPC. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] The Illumos Project
On Wed, Aug 4, 2010 at 10:24 AM, Paul Gress pgr...@optonline.net wrote: On 08/ 4/10 09:40 AM, Matthias Pfützner wrote: I wasn't refering to Illumos adding anything. I just mentioned, that nVIDIA does write the driver for all versions of Solaris and that they are therefore the focal point of contact, just like you mentioned. I only asked you, as you are in good contact with nVIDIA and might have known anything about their possible plans w.r.t. CUDA or OpenCL... ;-) But: I know: No Comment...;-) And: Yes, if there's business, they'll do it, that's for sure... Yes I believe that would be true if the OS was developed for Workstations. The only commitment I see from Oracle is Servers and Databases. Would CUDA help there? CUDA and OpenCL seem to be heavily used in HPC for general purpose calculations - not necessarily for displaying to a screen. For example... http://www.nvidia.com/object/io_1226945999108.html -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Community distro
On Sat, Jul 24, 2010 at 12:07 PM, Joerg Schilling joerg.schill...@fokus.fraunhofer.de wrote: Alan Coopersmith alan.coopersm...@oracle.com wrote: Can Oracle Engineering provide someone like Joerg what he needs to update his external SchilliX distro to implement ON snv_145/IPS 134? I'm guessing you're talking to me? I have no idea what Joerg needs, but we've already provided the community with everything they need to build current ON IPS bits, as evidenced by multiple external community members doing so. You do need a recent build of IPS to build ON builds 136 later, and X builds 144 later, but that's fully available from the pkg-gate sources on the opensolaris hg repos. Do you have a IPS SVr4 package or do you have a IPS source that compiles on a machine without IPS? Jörg Some time ago (about the B130 era...) I'm pretty sure it built SVr4 packages. An existing installation is not required to build it. $ hg clone http://hg.genunix.org/pkg-gate.hg/ $ cd pkg-gate.hg/src $ export PATH=/opt/sunstudio12.1/bin:$PATH : fix a bunch of .po files with mismatched newlines : http://cr.opensolaris.org/~mgerdts/pkg-po.patch.gz $ dmake ... /usr/bin/python2.6 setup.py build running build running build_py running build_ext $ echo $? 0 $ uname -sr Linux 2.6.34-12-default It is intended to be very cross platform and is the underlying update mechanism in Oracle products other than OpenSolaris. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Location of ZFS Uberblock
On Wed, Jul 7, 2010 at 2:19 PM, Mike Toms mjack...@unitrends.com wrote: Hello, My setup is a single disk with ZFS on the root fs. With UFS the superblock is 8KB from the beginning of the disk. Is there a way for me to find how much this offset is in ZFS? I have tried 'od -A x-x c1t2d0s0 | grep 00ba b10c' , but it does not find any matches. You are having a bit of od troubles... # mkfile -n 100m /tmp/100m # zpool create junk /tmp/100m # od -X /tmp/100m | grep 00bab10c 041 00bab10c 0016 0412000 00bab10c 0016 0414000 00bab10c 0016 141 00bab10c 0016 1412000 00bab10c 0016 1414000 00bab10c 0016 61641 00bab10c 0016 616412000 00bab10c 0016 616414000 00bab10c 0016 61741 00bab10c 0016 617412000 00bab10c 0016 617414000 00bab10c 0016 -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] New to OSOL: Move from Linux/Xen: BrandZ, xVM, virtualBox?
On Sat, Jun 19, 2010 at 4:46 AM, Robert Milkowski mi...@task.gda.pl wrote: On 19/06/2010 01:26, Mike Gerdts wrote: On Fri, Jun 18, 2010 at 6:33 PM, Dennylinuxm...@4lin.net wrote: hi, I want to move from a Linux (Debian Lenny) based RootServer (Hetzner) to OSOL. At the moment my Box runs under Xen and Debian Lenny with four virtual hosts. Now I don't know what is better. Web, Mail and Jabber are own virtual hosts. All runs with standard applications, like Postfix, Apache2, ejabberd so, what es the better way. Should I stay on Xen, or get i better performance with zones (brandz, for Linux if I need) ? Any suggestions? Brandz to run Linux would be a bad idea because it was recently end of featured and removed from the source. http://hg.genunix.org/onnv-gate.hg/rev/4c5722bc28dc Supposedly there is going to be a new release of OpenSolaris in the first half of 2010. Oracle ends their quiet period before announcing financial results on June 24. Given the quiet period and the first half commitment, that means that in 7 - 12 days there is a high likelihood that there will be a new OpenSolaris release. It would probably be smart to see what that release contains before making a decision on Xen vs. VirtualBox. I think brandz/Linux support will still be there as it was removed past snv_134. Nevertheless I wouldn't deploy brandz/linux as it is a dead end. That was my point. Also, as others have said, where you can get things to work natively under OpenSolaris, zones are going to be the most efficient. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] New to OSOL: Move from Linux/Xen: BrandZ, xVM, virtualBox?
On Fri, Jun 18, 2010 at 6:33 PM, Denny linuxm...@4lin.net wrote: hi, I want to move from a Linux (Debian Lenny) based RootServer (Hetzner) to OSOL. At the moment my Box runs under Xen and Debian Lenny with four virtual hosts. Now I don't know what is better. Web, Mail and Jabber are own virtual hosts. All runs with standard applications, like Postfix, Apache2, ejabberd so, what es the better way. Should I stay on Xen, or get i better performance with zones (brandz, for Linux if I need) ? Any suggestions? Brandz to run Linux would be a bad idea because it was recently end of featured and removed from the source. http://hg.genunix.org/onnv-gate.hg/rev/4c5722bc28dc Supposedly there is going to be a new release of OpenSolaris in the first half of 2010. Oracle ends their quiet period before announcing financial results on June 24. Given the quiet period and the first half commitment, that means that in 7 - 12 days there is a high likelihood that there will be a new OpenSolaris release. It would probably be smart to see what that release contains before making a decision on Xen vs. VirtualBox. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] PC Emulator for OpenSolaris?
On Mon, Jun 7, 2010 at 5:43 PM, Andrew Greimann agreim...@live.com wrote: I'm aware of the 'GPL situation'. As stated in my previous post yesterday, Apparently, as much as I like using OpenSolaris, without the ability to completely change the GPL components to where they're dwindled to simply coreutils and a LILO bootloader or something miniature in respect to the whole of the OS... It would be impossible to take out core components, but I believe avoiding using the GPL can be helped. I'm aware a total 100% is impossible, but a 80-85% non-GPL open source system I think is more realistic. Even Haiku under the MIT uses coreutils, and OS X uses at least nano under its Unix subsystem--so it's not [i]totally[/i] BSD. - OK Getting back to the post. After looking through Google, Win4Solaris seems to be the best solution over VirtualBox or JPC. I remembered the icon, but not the name. I merely started this discussion because I couldn't remember the name 'VirtualBox'. :-D Honestly, I had no idea it was GPLed--until seeing the webpage, I thought it was under a proprietary license. Is the LGPL as evil as the GPL in your mind? If so, sounds like you oughta steer clear of Win4Solaris as well. According to http://en.wikipedia.org/wiki/QEMU, Win4* are based on QEMU, which at least partially falls under the LGPL. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Dev repository last catalog update is March 6, 2010?
On Thu, Jun 3, 2010 at 12:42 PM, Harry Putnam rea...@newsguy.com wrote: Mike Gerdts mger...@gmail.com writes: On Sun, May 30, 2010 at 3:09 AM, Erik Trimble wrote: All that said, I'm still a little mystified as to why the normal development builds are being held back. Could it be because pkg image update is known to work pretty well going forward (b134 - b134b) but is known not to work well or is untested for going backward (b142 - b134b)? Or could there be other things (e.g zfs version 23 in b135) that would make going back to b134b problematic? If so, I suspect that this is a matter of protecting people from getting into a state where they can't transition from a dev build to a release build. People that really want to do development can (subject to periodic hiccups) do development on current bits by building their own. Sorry to side track a little here... I've seen that mentioned in several places recently, about building from sources. I wondered if there is a cache of info about doing that somewhere? Maybe some basic instructions or general outline of how to go at it? Good starting points are: http://hub.opensolaris.org/bin/view/Project+indiana/building_on http://hub.opensolaris.org/bin/view/Main/downloads#HBuildingOpenSolaris When you stumble trying to follow these processes (some of which will be slightly out of date at any given time) it is best to ask for help at tools-discuss, on-discuss, or another list where the developers are more likely to hang out. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Dev repository last catalog update is March 6, 2010?
On Sun, May 30, 2010 at 3:09 AM, Erik Trimble erik.trim...@oracle.com wrote: All that said, I'm still a little mystified as to why the normal development builds are being held back. Could it be because pkg image update is known to work pretty well going forward (b134 - b134b) but is known not to work well or is untested for going backward (b142 - b134b)? Or could there be other things (e.g zfs version 23 in b135) that would make going back to b134b problematic? If so, I suspect that this is a matter of protecting people from getting into a state where they can't transition from a dev build to a release build. People that really want to do development can (subject to periodic hiccups) do development on current bits by building their own. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Why are prices for Solaris 10 OS support not on Oracles Web pages?
On Sun, Apr 18, 2010 at 3:24 AM, Edward Martinez mindbende...@live.com wrote: I think since many business are moving away from HP-UX , Solaris ,etc over to redhat linux and windows, .Oracle, once they finish intergrating SUN, will offer solaris to the low-end business on sun units to compete against IBM, redhat, microsoft directly and OpenSolaris will continue like Mr. Roberts stated it will, will end up being the equivalent to Fedora. http://www.oracle.com/us/support/systems/premier/index.html says: Operating System, Firmware and Embedded Software updates * for Oracle Solaris * for Oracle Enterprise Linux * for Oracle VM * for your Oracle hardware My read of this is that, the Premier Support for Systems appears to get you support for Solaris, Oracle Enterprise Linux, and Oracle VM. If you would have otherwise paid for high-end support for Solaris, RHEL, and ESX on a box this may be attractive if you are willing to drink all Oracle koolaid. Of course, this depends on the cost of the system you are buying and your typical discount with the various software vendors. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Save OpenSolaris (according to Katonda)
On Fri, Apr 2, 2010 at 2:22 AM, Peter Jones bloosk...@netscape.net wrote: April fool or not..there is a serious side here.Is the uncertainty hitting user adoption rates or churn? Would it be a good idea to simply look at a mid term programe this year to rewrite or replace many of the non opensource packages to close the gap and silence critics? Maybe this is just good house keeping ..pity it takes away resources from other development innovations? Clearly I read into this a confusion on the publics part between Solaris and Opensolaris brands which comes down to marketing weakness? There already is such a project. http://hub.opensolaris.org/bin/view/Project+emancipation/Helping In addition, the ksh93 integration project has delivered (and has a lot more in the queue) a lot to help this goal. http://hub.opensolaris.org/bin/view/Project+ksh93-integration/ There are other efforts that have been ongoing for some time (e.g. ON build with gcc) that are also related. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Usage of psinfo_t from procfs.h
On Sun, Mar 14, 2010 at 6:01 AM, Ivan Petrov free...@abv.bg wrote: Hi, I'm trying to write a small piece of code that should be capable of getting the command line arguments for a given process. Similar to the ps(1) tool. What I can't figure out is how exactly to use the pr_argv field of the psinfo_t struct... Or maybe more similar to pargs? http://src.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/cmd/ptools/pargs/pargs.c -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] about zfs exported on nfs
On Mon, Mar 8, 2010 at 10:54 PM, Harry Putnam rea...@newsguy.com wrote: summary: A zfs fs set with smb and nfs on, and set chmod g-s (set-gid) with a local users uid:gid is being mounted by a remote linux host (and windows hosts, but not discussing that here). The remote user is the same as the local user in both numeric UID and numeric GID The zfs nfs/cifs share is mounted like this on a linux client: mount -t nfs -o users,exec,dev,suid Any files/directories create by the linux user end up with nobody:nobody uid:gid and any attempt to change that from the client host fails, even if done as root. It looks to me like you are using NFSv4 and the NFS mapping domains do not match. See /etc/default/nfs on OpenSolaris and I don't know what on Linux. Details: I'm not sure when this trouble started... its been a while, long enough to have existed over a couple of builds (b129 b133). But was not always a problem. I jumped from 129 to 133 so don't know about builds in between. I have a zfs_fs .. /projects on zpool z3 this is a hierarchy that is fairly deep but only the top level is zfs. (Aside: That is something I intend to change soon) That is, the whole thing, of course, is zfs, but the lower levels have been created by whatever remote host was working there. z3/projects has these two settings: z3/projects sharenfs on z3/projects sharesmb name=projects So both cifs and nfs are turned on making the zfs host both a zfs and nfs server. Also when z3/projects was created, it was set: chmod g-s (set gid) right away. The remote linux user in this discussion has the same numeric UID and GID as the local zfs user who is owner of /projects Later, and more than once by now, I've run this command from the zfs host: /bin/chmod -R A=everyone@:full_set:fd:allow /projects to get read/write to work when working from windows hosts. The filesystem is primarily accessed as an nsf mounted filesystem on a linux (gentoo linux) host. But is also used over cifs by a couple of windows hosts. On the linux client host, `/projects' gets mounted like this: mount -t nfs -o users,exec,dev,suid That has been the case both before having the problem and now. The trouble I see is that all files get created with: nobody:nobody as UID:GID, even though /projects is set as normal USER:GROUP of a user on the zfs/nfs server. From the remote (we only deal with the linux remote here) any attempt to change uid:gid fails, even if done by root on the remote. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] job interview help!
On Fri, Mar 12, 2010 at 8:36 AM, Mike DeMarco mikej...@yahoo.com wrote: everything else on the SAN. Don't forget to minimize and harden the build as much as possible before you patch, and patch (including firmware) before you let users of any type on it. Can you explain the logic behind minimize and harden before you patch? I have always fully patched them JASS then minimize. Hardening should always be verified after patching. Solaris patches quite commonly whack hardening applied to sendmail. I would think that if I minimized first it would save some time patching but that something could get missed in the patch install. Something like if one package is not installed a patch to a library that is used by another packages could get missed. If your minimization is such that patching breaks, the order doesn't matter. At some point x months or years in the future you will need to patch again. Don't minimize to the point that patching breaks. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Contributing code?
2010/2/8 Shawn Walker swal...@opensolaris.org: On 02/ 8/10 03:47 PM, Thommy M. Malmström wrote: After attending FOSDEM this weekend I'm interested in contributing to the OpenSolaris project. I'm not a brilliant coder, I'm not great with the logic of code and I've only been taught Java. But I'm not bad at formatting code and knowing where comments should and shouldn't be. Is code fairly clean already? Could I use my free time to reformat code to conform to a specific style? If so how can I download the code for OpenSolaris and make these changes? Or would it require me to pick a project within OpenSolaris (JDS/Kernel/IPS/etc)? Not to disappoint you, but the code has been worked on for +20 years, so it's fairly clean I would say. There are plenty of opportunities in this area still, in part because ti is 20+ years old :) Such as last: http://src.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/cmd/last/last.c I think this code looks good only with :set ts=2 -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] What's Going to Happen to OpenSolaris?
On Thu, Jan 28, 2010 at 10:20 PM, Gary Bainbridge g_patri...@yahoo.com wrote: If I wanted to run a GNU/Linux distribution I would, but apparently the decision is being made for those who like Solaris, to be made to run another Linux-type server. Seriously, how long (how many years) and how much money is it going to take to make OpenSolaris a replacement for Solaris 10? Is Oracle going to spend that much money? In my mind the key things that seem to be missing for it to be able to take the baton from Solaris 10 are: - Interactive text installer (apparently that came out in the last day or so) - Easy to configure automated installer. Some docs to go with the bootable AI iso would be a big help here. - The ability to host all of the media required on my own install servers - The ability to install software from a file - Speed improvements for pkg command, particularly on CMT - Package signing to ensure integrity of bits delivered. - Refactor package names - Integration with 800-USA-4SUN. Any OpenSolaris cases I have opened on a supported system have turned into redirects to mailing lists. Maybe this has improved in the past 6 months. - Stablization freeze and beta cycle Certainly, there may be other things that are missing from a Solaris 11 roadmap that no one has shared with me. However, those aren't necessarily things that keep it from being at least as good as Solaris 10. I suspect that if resources were increased to finish off the areas listed above that the beta cycle could begin with 3 - 6 months and a release could happen this year. By and large, I find that (after installation) OpenSolaris servers are nearly identical to Solaris 10 servers to administer, except when I need a feature that is unique to OpenSolaris. I generally find more value in the OpenSolaris feature set. Sure there are bleeding edge issues that crop up, but those are the things that get ironed out in stablization and beta. To make a desktop OS work as a data center OS is not remotely the best engineering practice. Could you run Solaris 8 on a desktop? Sure. But why? It wasn't practical. Could you use Windows 95 as a server? Probably many did. But why? That wasn't its intended use. Now the OS is going to be retrofitted to make it an enterprise server? With Solaris you can choose what you want to install. Not so with OpenSolaris. You get what you're told. I see it just the opposite. On Solaris 10 if you are trying to create an image to work across all hardware platforms, the most supported way of doing this is to install SUNWCXall. Then you can start trimming away big things like Staroffice and the bulk of the desktop tools until you become afraid that you are going to break some application's unknown dependencies. For example, the Oracle installer never says that it needs to have X libraries (but it does if running in GUI mode) and some annoying J2EE app never mentions that it needs X fonts but it does else it can't render some text in images. Go figure. If you decide not to do flash archives on Solaris 10 and install via pkgadd instead (not nearly as painful after Casper's work on turbo charging), you can start out a lot smaller (e.g SUNWCmreq or SUNWCreq), but creation of a custom profile to bring in those annoying bits that you've learned are needed is quite a pain. If you guess wrong at install time, you will go through lots of iterations of pkgadd + error messages to try to add the other packages you need - once you figure out which uninstalled package delivers the shared library or font files that you need. In contrast, with OpenSolaris I can use AI to start out with a pretty small installation. Lots smaller than slim_install (live CD contents) - look to the packages installed in a zone by default (plus kernel bits) as a rough guide of the base install. Then add the packages that deliver the shared library or font that you need. The dependencies should automatically add the rest of the packages. I know this isn't perfect yet, but I think a beta cycle would raise the bugs and fixes for missing dependencies. If you miss something at install time, it is just as easy after the system is installed to add the required packages with automatic dependency resolution. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] What's Going to Happen to OpenSolaris?
On Wed, Jan 27, 2010 at 9:49 PM, W. Wayne Liauh w...@hawaiilinux.us wrote: While OpenSolaris is already open source (mostly), there were some plans within Sun to open source more technologies, but Sun didn't get to it because of the merger. In particular - Sun Studio, C++ compiler etc. Interesting what are Oracle intentions in regards to keeping those plans? I believe the main issue is not--in case the unthinkable happens--whether Oracle will open source more technologies, but, as another poster mentioned, will Oracle continue to license update/upgrade of the existing code under the CDDL license. All of Sun's open-sourced code is under dual license, meaning that Oracle can continue to develop Sun's technologies under a proprietary license, but freeze the CDDL'd code at its current state. Of course, anyone can fork the Solaris kernel under the CDDL scheme. But the possibility is zero. Sun has already been doing that. If you pay for support and use the support repository, you are using a closed fork. Since OpenSolaris-dev and OpenSolaris-support are pretty close together temporally, this isn't a big deal. You can look at a snapshot of the snv_111 code and the list of fixes applied to get a pretty good idea of what the code looks like. After several years and potentially thousands of fixes, a lot of the benefit of the open source roots is lost. While the typical customer doesn't have an interest in modifying the code, many have an interest in looking at it to understand observed behavior or to aid in writing dtrace scripts that journey into fbt probes. As the years have passed since the fork between what became Solaris 10 and what became OpenSolaris, I have increasingly less confidence that looking at any version of OpenSolaris code will allow me to really understand what is happening on a Solaris 10 system. That is, as the number of fixes and features included in Solaris 10 increases, the value of the open source roots decreases. I have always expected that the same will happen with Solaris 10+1 (11g?). I have consistently asked Sun to make the code for supported OS's available to customers, even if it is under a license other than the CDDL. I encourage others to make similar requests. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Make Opensolaris boot failsafe
On Sat, Jan 16, 2010 at 1:53 PM, Will Meister meister.w...@gmail.com wrote: I maintain several opensolaris fileservers remotely. Occasionally, one will fail to boot requiring an on-site visit to fix. It's usually something minor, like a service failing to start following an upgrade or a corrupt boot archive. For example, editing /etc/system or another config file prior to the system losing power or crashing will cause the system to fail to boot on restart as it thinks the boot archive is corrupt. (It is apparently only regenerated during a successful shutdown/reboot). Is there a way to either force the system to boot to console even if something is awry, or at least provide SSH access to the recovery console? The servers in question don't have IPMI and I'd prefer to avoid having to get an IPKVM for each one. Any ideas or suggestions would be appreciated. You could probably configure the systems to use a serial console instead of a KVM console. At that point, you could either use something like a cyclades terminal server (still somewhat costly) or cross-link pairs of servers. This assumes that each server has two serial ports - put the console on serial port 0 and connect it to serial port 1 of its buddy server with a null modem cable. You can use tip(1M) or other things that probably aren't included in OpenSolaris to talk to the serial port. But really... you should be working toward buying enterprise class servers that have a built-in system controller so that you don't have to worry about this stuff. If you don't have enterprise class servers, I suspect that you are also missing things like ECC memory and redundant power supplies. You can run a small shop like this for a while, but eventually the reasons to increase the quality of your hardware will reach a tipping point. Something like a slightly bad DIMM can make your life miserable with corrupted data (possibly causing more frequent reboots that don't go well...) and unsympathetic ears on zfs-discuss. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] OpenSolaris, Solaris 10, SXCE, Indiana: which to use for a developer?
On Sat, Jan 2, 2010 at 1:50 PM, Ali Bahrami ali.bahr...@sun.com wrote: Jürgen Keil wrote: On Sat, Jan 02, 2010, Jürgen Keil wrote: [OpenSolaris Indiana: GNU?] Well, it does have both the gnu utilities (/usr/gnu/bin) and the solaris counterparts, and in the default setup the /usr/gnu/bin directory is ahead of /usr/bin in $PATH. I simply removed /usr/gnu/bin from $PATH, but I still got vim instead of vi[1] (in /usr/bin AFAIR), Yep, that's true. OpenSolaris does not have the real vi. Not sure what the reason is; the source seems be available in the onnv mercurial repository, so I guess the real vi would be re-distributable with OpenSolaris. and bash as root sh. Ok, so change root's shell in /etc/passwd. If there is a way to configure OpenSolaris to be(have) like Sun's Solaris (for which I develop software and to which I'm used (besides *BSD)), then please let me know. If you target Solaris 10, then you should probably run Solaris 10 on your development system, too. The old vi and Bourne shell (sh) are in /usr/has/bin. - Ali And next you are likely to discover that the default installation makes for a fairly functional desktop but is worthless for development. To get a full installation (ala SUNWCXall in previous releases): # pkg install redistributable As someone else said, if you are trying to write software to run on Solaris 10, I would highly recommend developing on Solaris 10 so that your binaries are linked against libraries that work on Solaris 10 and OpenSolaris. If you produce binaries on OpenSolaris, they will not run on Solaris 10. Your aversion to a GNU userland (which is fine) somehow makes me think that OpenSolaris will have few benefits for you over Solaris 10. There are certainly lots of benefits to OpenSolaris, but unless your program or development environment will take advantage of them you will probably be more productive on Solaris 10 until OpenSolaris matures a bit. If you are trying to build things that require recent versions of various open source libraries, you will probably have a much easier time on OpenSolaris. This is because the freeware that came with Solaris 10 was not necessarily at the latest versions by the time that Solaris 10 was first released 5 years ago (wow, it's been 5 years) and I don't think that there has been a systematic effort to refresh them since then. When security issues or bugs came up, it is probably more likely that the old versions were fixed rather than transitioning to new versions. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Building ON on opensolaris
On Sun, Dec 27, 2009 at 10:13 AM, Vikram Hegde vikram.he...@sun.com wrote: Hi, Has anything changed with building ON ? I followed the current instructions on the ON community web-page and had to jump through hoops to compile the current onnv-clone on OSOL-build 130. This included installing a bunch of missing packages such as SUNWapch SUNWprd etc. I started out by installing entire but that didn't drag in the above packages. Are the ON build instructions up to date ? The entire package is an incorporation that (mostly? only?) establishes compatibility rules. If you want to install everything what you really want is: pkg install redistributable But that doesn't really get you what you need either. This is because there are some things required to build ON (e.g. WBEM) that is not redistributable. Also, the version of the Studio compilers that are in the repository have not yet been updated to the right patch level. See the following for the instructions you need. Do not skip the extras repository, as that is where the non-redistributable WBEM is found. http://hub.opensolaris.org/bin/view/Project+indiana/building_on http://hub.opensolaris.org/bin/view/Community+Group+tools/sun_studio_12_tools And then send mail to on-discuss and/or tools-discuss pointing to the page that led you to the unhelpful documentation. FWIW, those lists will tend to have people more well-versed in building ON than opensolaris-discuss. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] build software in a non global ipkg-zone
On Sun, Dec 20, 2009 at 2:55 PM, Dr. Martin Mundschenk m.mundsch...@mundschenk.de wrote: Hi! I'd like to build netatalk in a non global zone. I installed gcc-dev in the ipkg-zone, but make stops with the error make: Fatal error: Command failed for target `all-recursive' I guess, some libraries or executables are missing. How can I install an ipkg-zone such that all crucial elements are available to build software packages from source? Regards, Martin Posting more detailed error messages would be helpful. My guess is that you will see some message about a missing header (*.h) or library. If the build process complains about a missing header file, use pkg search whatever.h to find the name of the package to install. If it complains about not being able to find a library for -lfoo, you should search for libfoo.so. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Syslog-ng?
On Mon, Dec 7, 2009 at 8:34 AM, Eugene Vilensky evilen...@gmail.com wrote: Hi, Is syslog-ng available for OpenSolaris, through IPS or otherwise? If I were to build it, is there an easy metapackage for the build system, maybe analogous to Red Hat's rpmbuild+dependencies or Debian's build-essential metapackage? I doubt that if it were installed on OpenSolaris that it would work for processing logs for local programs. This is because the things that do the logging are going to use libc's implementation of syslog(3C), which logs via a door (see libdoor(3lib)). Doors are unique to [Open]Solaris and as such syslog-ng is unlikely to be listening for log messages at /etc/.syslog_door. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] crontab
On Wed, Dec 2, 2009 at 10:29 PM, Edward Ned Harvey sola...@nedharvey.com wrote: I searched the OpenSolaris Bible and the internet but didn't find a suitable answer to how to configure cron jobs in OpenSolaris? Any hint? Personally I don't like crontab -e because the editor variable needs to be set. Also there's no backup if you make a mistake. Instead, I like to use crontab -l .crontab and then vi .crontab and then crontab .crontab This was fixed three years ago (snv_39, S10u5). http://hg.genunix.org/onnv-gate.hg/rev/e8dfcbd4692e http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6344121 -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] How do I install patches for sun studio 12 update 1 for opensolaris?
On Sun, Nov 29, 2009 at 4:39 PM, Shawn Walker swal...@opensolaris.org wrote: keithk wrote: Hi: I am getting some compilation error which I think one of the patches listed on http://developers.sun.com/sunstudio/downloads/patches/ss12u1_patches.jsp for Sun Studio 12 update 1 would fix the problem. I downloaded the patches, but could not patch the software because patchadd is not part of opensolaris. How do I patch Sun Studio in Opensolaris? You don't. The only way to get updated versions of packaged software in OpenSolaris is using the package system. pfexec pkg install sunstudio12u1 ...will install the newest version of sunstudio12 update 1 available, or update the already installed version. It was last updated on June 17, 2009 at 10:54:13 PM CST. As of last week the versions of Sun Studio (12u1, express) available through the dev repository can not build ~b128 ON because it gets stuck trying to build ksh93 components[1]. 1. http://opensolaris.org/jive/thread.jspa?threadID=116483#437006 I ran into many other problems with C++ code as well. If memory serves, this thread[2] discusses the type of problems I saw. 2. http://opensolaris.org/jive/thread.jspa?threadID=110262 With slight modifications[3] I was able to get a good build with the Sept 2009 CBE[4]. 3. http://opensolaris.org/jive/thread.jspa?threadID=118399tstart=0 4. http://hub.opensolaris.org/bin/view/Community+Group+tools/sun_studio_12_tools It looks like there might be an effort[5] with the Use Case Project to fix this up. Rich may appreciate a little help with this. 5. http://mail.opensolaris.org/pipermail/usecase-discuss/2009-November/51.html -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] freezing a zone
2009/11/18 Gaëtan Lehmann gaetan.lehm...@jouy.inra.fr:
Hi,
Is it possible to freeze a zone and unfreeze it a few seconds later?
I can't find anything like that in the doc. In the mean time, I've done
that:
running_processes=`ps -o s= -o pid= -z $zone_name | grep -v T | awk '{print
$2}'`
kill -s SIGSTOP $running_processes
... do important stuff here...
kill -s SIGCONT $running_processes
You could probably use a dtrace script along the lines of the
following in the global zone...
# dtrace -wqn 'sched:::on-cpu /zonename == $1/ { printf(Stopping pid
%d/%d in %s\n, pid, tid, zonename); stop(); }' $zone
When you are ready to unfreeze it:
# prun `pgrep -z $zone`
--
Mike Gerdts
http://mgerdts.blogspot.com/
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Is there a driver for Marvell?
On Thu, Oct 8, 2009 at 8:49 PM, Larry Robinson lark...@aol.com wrote: I am using Solaris on one machine and looking at using opensolaris on another, just can't get enough of this Sun stuff. The reason I do not have opensolaris is the fact that I could not find a driver for my ethernet. It is a Marvell 88E8056 PCI-E gigabyte Rev(13). Does anyone know if there is a driver for opensolaris and marvell? -- It looks like there was a relevant putback a few weeks ago... http://hg.genunix.org/onnv-gate.hg/rev/fa1b0b185963 The tag for build 124 of ON came 3 days later... http://hg.genunix.org/onnv-gate.hg/tags And the latest in the dev branch seems to be build 124... http://pkg.opensolaris.org/dev/info/0/entire%400.5.11%2C5.11-0.124%3A20090925T222419Z -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] /dev/null: how to revive?
On Fri, Oct 2, 2009 at 8:21 AM, Sh ssl_hun...@yahoo.com wrote: Vikram Hegde: i can't do anything with /dev/null Not with ln or some special tools like devlinks or devfsadm I can't do even touch /dev/null This is not just a problem of removal /dev/null, this a serious bug My problem is similar to http://www.opensolaris.org/jive/thread.jspa?messageID=271702#271702 Note that /dev in OpenSolaris is dev(7fs) not zfs or ufs as it was in S10. I suspect that this is affecting the normal approaches for fixing /dev/null. It is probably best to open a bug on this. Folks over at on-discuss may be better able to help with a workaround. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] rsh ?
On Fri, Oct 2, 2009 at 1:07 PM, Jim Langston jim.langs...@sun.com wrote: Hi all, I need to have rsh access across my grid, I have set up access on Solaris, Fedora and RHat, but on my OpenSolaris system I am not able to execute an rsh command without getting a connection refused. I can login w/o a passwd (which is what I have been setting up). I have tried to execute locally: langs...@zulusolaris:~# rsh zuluSolaris date connect to address :::192.168.1.206: Connection refused Trying ::1... zuluSolaris: Connection refused Executing from the remote Solaris system: -bash-3.00$ rsh zuluSolaris date zuluSolaris: Connection refused -bash-3.00$ But, I can rsh just fine, as long as I don't try the command: langs...@zulusolaris:~$ rsh zuluSolaris Last login: Fri Oct 2 13:36:55 from zuluSolaris Sun Microsystems Inc. SunOS 5.11 snv_111a November 2008 -bash-3.2$ rsh with no command really uses rlogin. This means that zuluSolaris has svc:/network/login:rlogin online but svc:/network/shell:default is not. Or, from the remote Solaris system: -bash-3.00$ rsh zuluSolaris Last login: Fri Oct 2 14:04:14 from zuluSolaris Sun Microsystems Inc. SunOS 5.11 snv_111a November 2008 -bash-3.2$ Any hints/thoughts ? Thanks, Jim -- / Jim Langston Sun Microsystems, Inc. (513) 702-4741 (Cell) (877) 854-5583 (AccessLine) AIM: jl9594 jim.langs...@sun.com ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SSH with public keys not working (not recognizing the key file)?
On Wed, Sep 23, 2009 at 6:31 PM, Jonathan C. Bailey jbai...@co.marshall.ia.us wrote: Everything is on one line.. Actually, the authorized_keys file was copied (working) from an Ubuntu 8.04 system.. Here's an example below. I've removed the actual key text in this case to shorten up the line (but everything *is* on one line)... command=/backup/validate-rsync ntso ssh-rsa KEY_TEXT rsyncbac...@ntso -Jon What are the permissions on the authorized_keys file and all the directories leading up to it? 1. Can the user read the file? 2. Can anyone else write to the file or any parent/ancestor directory? -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SSH with public keys not working (not recognizing the key file)?
On Wed, Sep 23, 2009 at 8:31 PM, Jonathan C. Bailey jbai...@co.marshall.ia.us wrote: / and /home have correct permissions. rsyncbackup is another UID 0 user (don't ask - it's worked fine on other platforms). All the SSH related permissions seem fine too. I've also tried authorized_keys and authorized_keys2. I've copied the same keys to root and it still works. Solaris ssh is a hybrid of OpenSSH and Solarisisms. Your configuration may be tripping over a Solarisism. Perhaps leading with this at security-discuss will get the attention of someone more familiar with the details of the differences. They response will probably include something like: Why not use a non-UID 0 account with RBAC and have the following in authorized_keys? command=/usr/bin/pfexec /backup/validate-rsync ntso ssh-rsa KEY_TEXT rsyncbac...@ntso BTW, I *did* turn root into a real user and added the 'PermitRootLogin yes' line to sshd_config. -Jon - Original Message - From: Mike Gerdts mger...@gmail.com To: Jonathan C. Bailey jbai...@co.marshall.ia.us Sent: Wednesday, September 23, 2009 8:25:20 PM GMT -05:00 Colombia Subject: Re: [osol-discuss] SSH with public keys not working (not recognizing the key file)? On Wed, Sep 23, 2009 at 7:13 PM, Jonathan C. Bailey jbai...@co.marshall.ia.us wrote: /home/rsyncbackup is 700 /home/rsyncbackup/.ssh is 700 /home/rsyncbackup/.ssh/authorized_keys2 is 600 I assume that the user rsyncbackup is also the owner of all of those. (grasping at straws a bit...) And how about / and /home? They should be writable by only root. sshd is also picky about permissions on the path to the sshd_confg file. Check the permissions on /etc, /etc/ssh, and /etc/ssh/sshd_config. Also removed the command option from the key (so it started with ssh-rsa), and no difference.. Same problem... Here's the server side of that exchange: debug1: userauth-request for user rsyncbackup service ssh-connection method publickey debug1: attempt 1 initial attempt 0 failures 1 initial failures 0 debug2: input_userauth_request: try method publickey debug1: test whether pkalg/pkblob are acceptable debug1: temporarily_use_uid: 0/0 (e=0/0) debug1: trying public key file /home/rsyncbackup/.ssh/authorized_keys debug1: restore_uid: 0/0 debug1: temporarily_use_uid: 0/0 (e=0/0) debug1: trying public key file /home/rsyncbackup/.ssh/authorized_keys2 I see here that it is looking at authorized_keys and authorized_keys2. Which file did you add it to? What happens if you try the other? debug3: secure_filename: checking '/home/rsyncbackup/.ssh' debug3: secure_filename: checking '/home/rsyncbackup' debug3: secure_filename: terminating check at '/home/rsyncbackup' debug1: restore_uid: 0/0 debug2: key not found debug2: userauth_pubkey: authenticated 0 pkalg ssh-rsa Failed publickey for rsyncbackup from 10.243.193.3 port 52043 ssh2 I seriously doubt this is your problem, but usernames 8 characters long are not supported. -Jon - Original Message - From: Mike Gerdts mger...@gmail.com To: Jonathan C. Bailey jbai...@co.marshall.ia.us Cc: opensolaris-discuss opensolaris-discuss@opensolaris.org Sent: Wednesday, September 23, 2009 6:57:16 PM GMT -05:00 Colombia Subject: Re: [osol-discuss] SSH with public keys not working (not recognizing the key file)? On Wed, Sep 23, 2009 at 6:31 PM, Jonathan C. Bailey jbai...@co.marshall.ia.us wrote: Everything is on one line.. Actually, the authorized_keys file was copied (working) from an Ubuntu 8.04 system.. Here's an example below. I've removed the actual key text in this case to shorten up the line (but everything *is* on one line)... command=/backup/validate-rsync ntso ssh-rsa KEY_TEXT rsyncbac...@ntso -Jon What are the permissions on the authorized_keys file and all the directories leading up to it? 1. Can the user read the file? 2. Can anyone else write to the file or any parent/ancestor directory? -- Mike Gerdts http://mgerdts.blogspot.com/ -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] ssh not picking up new project
On Tue, Sep 22, 2009 at 8:27 AM, Kicklighter, Nathan CTR USAF AMC AMC/A6IM nathan.kicklighter@scott.af.mil wrote: None that I know of. I will try coping the pam.conf from SUNWcsr over my current and reboot once my user testing is complete. Thank you much for your help thus far. I will let you know how it goes. No need to reboot and in fact rebooting is not recommended. When mucking with /etc/pam.conf, be sure to keep a backup copy of one that works and never log out of your last root shell until you confirm you can get in as root (via console login, su, or whatever else makes sense in your environment). Another thing that may be biting you is using a third-party ssh implementation. The fact that you seem to have a customized pam.conf suggests that someone may be applying a security policy that was appropriate in the Solaris 8 days when Solaris didn't come with ssh. //SIGNED// Nathan Kicklighter, Contractor, AMC/A6IM DSN 576-4229 Comm (618) 256-2249 -Original Message- From: Mike Gerdts [mailto:mger...@gmail.com] Sent: Monday, September 21, 2009 5:45 PM To: Kicklighter, Nathan CTR USAF AMC AMC/A6IM; opensolaris-discuss@opensolaris.org Subject: Re: [osol-discuss] ssh not picking up new project On Mon, Sep 21, 2009 at 4:24 PM, Kicklighter, Nathan CTR USAF AMC AMC/A6IM nathan.kicklighter@scott.af.mil wrote: Sorry about that: sshd auth required pam_unix_auth.so.1 sshd account required pam_unix_auth.so.1 sshd password required pam_unix_auth.so.1 Some time back there was a re-whack of the PAM modules in Solaris. This caused the functionality of pam_unix_auth.so.1 to change and be split up between many other modules. It looks like your Solaris 10 system has been customized with a pam.conf that dates back to before this change (which I think happened late in Solaris 9 or early in Solaris 10). You probably need to either get rid of the lines for sshd (in which case they will inherit settings from other which may or may not be broken in your configuration) or update them so that pam_unix_cred.so.1 is called. You can find the default contents for pam.conf in /var/sadm/pkg/SUNWcsr/save/pspool/SUNWcsr/reloc/etc/pam.conf. Unless you need ssh to use some custom pam modules, you are best off not adding entries for sshd and instead letting it fall back to default. -- Mike Gerdts http://mgerdts.blogspot.com/ -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] ssh not picking up new project
On Mon, Sep 21, 2009 at 3:30 PM, camps support nathan.kicklighter@scott.af.mil wrote: We have created a new project for oracle 10g to grab some much needed memory settings. If I su to oracle. Is this su or su -? id -p uid=1001(oracle) gid=70(oinstall) projid=100(user.oracle) When I ssh to the server using oracle. id -p uid=1001(oracle) gid=70(oinstall) projid=0(system) This is resulting in my dba's needed to do a su before initiating anything with oracle and oracle starting at boot with the wrong systems settings. This is solaris 10 with oracle 10g. What does grep ssh /etc/pam.conf say? How about grep other /etc/pam.conf? It looks as though someone may have customized pam.conf and now something like the following line is being missed. other auth required pam_unix_cred.so.1 is not being called. (Substitute ssh for other if any lines say ssh auth ) Also, what are the permissions on /etc/project? -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] ssh not picking up new project
On Mon, Sep 21, 2009 at 4:24 PM, Kicklighter, Nathan CTR USAF AMC AMC/A6IM nathan.kicklighter@scott.af.mil wrote: Sorry about that: sshd auth required pam_unix_auth.so.1 sshd account required pam_unix_auth.so.1 sshd password required pam_unix_auth.so.1 Some time back there was a re-whack of the PAM modules in Solaris. This caused the functionality of pam_unix_auth.so.1 to change and be split up between many other modules. It looks like your Solaris 10 system has been customized with a pam.conf that dates back to before this change (which I think happened late in Solaris 9 or early in Solaris 10). You probably need to either get rid of the lines for sshd (in which case they will inherit settings from other which may or may not be broken in your configuration) or update them so that pam_unix_cred.so.1 is called. You can find the default contents for pam.conf in /var/sadm/pkg/SUNWcsr/save/pspool/SUNWcsr/reloc/etc/pam.conf. Unless you need ssh to use some custom pam modules, you are best off not adding entries for sshd and instead letting it fall back to default. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] [on-discuss] [pkg-discuss] ON/SXCE bi-weekly schedule not valid anymore?
On Tue, Aug 25, 2009 at 3:27 PM, johan...@sun.com wrote: On Tue, Aug 25, 2009 at 12:21:18PM +0100, Peter Tribble wrote: It's not just search that's slow - most pkg operations feel slow. How about providing some data instead of wild accusations? I can install entire, SUNWcs SUNWcsd, and babel_install all in under 7 minutes on a 100mbit network. Are you using modern hardware? Silly example, but is representative of something that needs to be done from time to time. Solaris 10: $ uname -srvi SunOS 5.10 Generic_141414-02 SUNW,SPARC-Enterprise-T5120 $ ptime grep -w ls /var/sadm/install/contents [snip] real2.771 user2.653 sys 0.115 OpenSolaris: $ uname -srvi SunOS 5.11 snv_111b SUNW,SPARC-Enterprise-T5120 $ ptime pkg search -l ls [snip] real 34.130866085 user 30.690454076 sys 0.774897050 My experience with hardware that I can order from Sun today says that the new software takes 12x longer for equivalent tasks. There is a very high startup cost with pkg that does not exist with pkgadd. With pkgadd I don't think too far ahead to be sure to group as many operations into one invocation as possible. When I use pkg, I most certainly try to lump as many operations as possible into each invocation to avoid this startup penalty. In balance, I believe that with a local repo, installation of large images will be faster with pkg than with pkgadd. Hopefully it is competitive with installation via flash archives. Even when upgrading across the internet, pkg image-update is way faster than luupgrade. When I look at the publicly disclosed/speculated road map for CMT systems, I don't see things improving for the simple operations without fixing the software. I eagerly await the SAT solver and any other improvements that are in the works. Right now I'm not complaining - I know the software is young and the primary development platform is x86 where the regression isn't so apparent. Once I start hearing that there aren't big performance improvements coming, I will start opening support calls if the performance is still worse than before. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] DRAFT2: OpenSolaris Use Case Project Proposal
On Thu, Aug 20, 2009 at 5:10 PM, Jim Walkerjames.wal...@sun.com wrote: Thanks for the great response. It looks like there is a need for this project :) I added the Systems Administration area. Technically, that is the Systems Administration Community Group. Just a few more details before we begin the project ratification process: - need someone to volunteer to be project point of contact Since you have taken the initiative to propose it, I nominate you. If there is a reason that you do not want this role, feel free to decline. - I would like one more Community to sponsor this project since it crosses Community Boundaries. If you think the project proposal can be ratified in your Community let me know. The Documentation and Testing Communities already plan to sponsor this project. I think it is appropriate for the Systems Administration Community Group (hereafter sysadmin community) to sponsor it as well. Peter Tribble (a core contributor in the sysadmin community) forwarded the proposal to sysadmin-discuss, suggesting he would support endorsement. I am also a core contributor in the sysadmin community and am supportive. BTW. Since the website transition is in progress this project may start by sharing a sponsoring Community's webpages and/or genunix.org. The usecase-discuss email list will be available immediately after the project is ratified. Cheers, Jim = o Project: OpenSolaris Use Case Project o Short Description: The OpenSolaris Use Case Project collects, develops and validates OpenSolaris operating system and application Use Cases and provides real world feedback to development, documentation, testing and other OpenSolaris groups to help ensure users get what they need from OpenSolaris distributions and applications. o Project Description: The primary focus of the OpenSolaris Use Case Project is to evaluate and organize real world OpenSolaris user experiences into Use Cases, case studies and best practices for the benefit of other OpenSolaris users, and to provide feedback to OpenSolaris developers, technical writers, testers and others so they may produce better quality products and education materials that more effectively meet user requirements. This project is run by users for users. Project roles and activities are diverse and include users bringing specific issues into the discussion, end users and consultants bringing expertise and focus to areas they are most interested in, Use Case organization and development experts, and project leaders who organize the overall project and review processes. Development, documentation, test teams and others can provide this project materials they want user feedback on prior to release to help ensure the user is getting what they need to be successful. Example areas of interest can include: * Core OpenSolaris Technologies zfs, smf, security, and DTrace and common OpenSolaris commands and utilities * OpenSolaris Installation and Packaging pkg, beadm commands, pkgmgr GUI, Automated Installer and Distro Constructor * OpenSolaris Systems Administration Standard and unique Systems Administration approaches and best practices * OpenSolaris Desktop OpenSolaris desktop, layout and appearance and standard desktop applications * Virtualization zones, crossbow, ldoms, xVM, VirtualBox, etc... * OpenSolaris Distributions distribution centric use cases for BeleniX, MilaX, Nexenta, SchilliX etc... * OpenSolaris Applications OpenOffice, contrib repo applications, Oracle, etc... * Development Tools SunStudio, netbeans, hg, svn, etc... * Hardware Configurations Various hardware types and storage configurations, etc... * Combination Solutions Unique End User solutions using multiple technologies * Use Case Development This area focuses on Use Case development itself, including how Use Cases are written, Use Case tools and the review process. Both formal and informal Use Case methods will be used. o Sponsoring Communities: Documentation Community Testing Community o Initial Project Members: I think this should be Project Members Joanie Diggs (joanied) Jason Schroeder (jks) Alex Smith (sentinel) Uros Nedic (urosn) Sriram Natarajan () Steven Acres (swa) Rich Reynolds (odd1) Octave Orgeron (oorgeron) Che Kristo (che) Guruprasad (lgp_cse) norbert (norbertc) Hernan Saltiel (hecsa) Masafumi Ohta (masafumi) Ken Mays (kmays) Mike Gerdts (mgerdts) o Point of Contact: (needed) o Project Needs: Standard project website and SCM setup at: http://www.opensolaris.org/os/project/usecase/ usecase-disc...@opensolaris.org email alias. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Migrating from SXDE/SXCE to OpenSolaris
and changing all occurrences of sxce_zfs to opensolaris. Now you should be able to reboot, select the opensolaris menu item and be good to go. Presumably you may need to copy a few items from the old boot environment - /etc/lu/synclist in the old boot environment will be a good place to start. At such a time as you are ready to whack the old UFS (and potentially swap), you can: - disable any swap that is not part of the zpool - add swap to the zpool if needed - detach one half of the mirrror, - repartition that half so that s0 has the whole disk, - reattach that disk, using s0 this time - run installboot? - detach the other half of the mirror - repartition - reattach - run installboot? - reboot - I think this is required to see the added space If you decide to go this route, please share your experience. I'm sure there will be some corrections to the procedure. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Python on Solaris Lacking Multiprocessing Support?
On Wed, Jul 29, 2009 at 10:16 AM, John Wythebitsp...@gmail.com wrote: Hello, I am trying to write a python application that uses the multiprocessing module on OpenSolaris snv_117. I am guessing the python packages for OpenSolaris do not ship with multiprocessing support, but I am lost as to why. I've searched the web extensively but can not find an explanation for this. Also, beware of the Global Interpreter Lock. http://raymondtay.blogspot.com/2008/11/python-performance-and-gil.html On a particular highly multi-threaded workload, I captured the following prstat -mL output below, notice that there is a lot of time waiting on user level locks. PID USERNAME USR SYS TRP TFL DFL LCK SLP LAT VCX ICX SCL SIG PROCESS/LWPID 16656 root 68 4.8 0.0 0.0 0.0 27 0.0 0.4 1K 0 13K 0 python2/152398 16656 root 68 3.2 0.1 0.0 0.0 17 12 0.3 1K 0 9K 0 python2/152397 16656 root 20 0.8 0.0 0.0 0.0 15 64 0.2 922 97 2K 0 python2/8 16656 root 8.8 1.4 0.1 0.0 0.0 51 39 0.5 2K 302 5K 0 python2/100558 16656 root 7.6 1.2 0.1 0.0 0.0 53 38 0.4 2K 303 4K 0 python2/53075 16656 root 7.4 1.2 0.0 0.0 0.0 53 38 0.4 2K 214 4K 0 python2/117545 Mike -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Python on Solaris Lacking Multiprocessing Support?
On Wed, Jul 29, 2009 at 12:07 PM, John Wythebitsp...@gmail.com wrote: Also, beware of the Global Interpreter Lock. http://raymondtay.blogspot.com/2008/11/python-performa nce-and-gil.html I only glanced over the article because I am aware of the issues around the GIL and threads. The multiprocessing module is a great work-around for this. http://docs.python.org/library/multiprocessing.html Thanks much for the clarification. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] OpenSolaris 0906 TIMEZONE
On Sun, Jul 12, 2009 at 2:39 PM, Gary Bainbridgegbainbri...@oriental.com wrote: The global zone has the correct timezone US/Central but all of my installed zones have a PDT date, even though /etc/default/init has US/Central. This must be a bug, and if so, how can I set my zones to US/Central? What does /etc/default/init in the non-global zones say? Each zone (really, each process) can have its own timezone. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] OVER 24 HOURS AND I HAVEN'T BEEN ABLE TO INSTALL A ZONE
On Fri, Jul 3, 2009 at 6:40 PM, Anon Y Moussystem5u...@yahoo.com wrote: What is? This message makes no sense on the mail list. Ian, I take it that you only use Solaris 10 and Solaris Express and don't ever use Indiana, ipkg zones or IPS package management? The main OpenSolaris IPS repository has been going down repeatedly for over two days now. Some people who have been running OpenSolaris on their servers have not been able to deploy a new ipkg zone in over 48 hours and all their work has been forced to grind to a halt. Nobody who works at Sun has said anything about what might be causing the problem. All the people who use ipkg zones on OpenSolaris are basically now S.O.L. because every time they try to deploy a new ipkg zone, the new ipkg zone automatically tries to download packages from the Sun repository and the download times out because the IPS repository is down and consequently the zone won't deploy. There are already three or four threads about this in the help list at opensolaris.org (see link below): http://www.opensolaris.org/jive/thread.jspa?threadID=107149tstart=0 but nobody is responding to the cries for help (just dead air time and crickets chirping). So we tried moving the thread to the discuss list in the hope that someone knowledgeable would see the multiple pleas for help and respond. Yesterday I received a reply from Shawn Walker. Not terribly informative, but at least it seems as though it is being worked. http://mail.opensolaris.org/pipermail/pkg-discuss/2009-July/014927.html A while later, there was a putback to the pkg source code repository that looks like it may be related. http://mail.opensolaris.org/pipermail/pkg-discuss/2009-July/014937.html -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Cleaning up /var/pkg/pkg directory?
On Sat, Jun 6, 2009 at 5:42 AM, Josip Gracin jgra...@inge-mark.hr wrote: Hi! Is there a way to clean up /var/pkg/pkg directory used by pkg(1)? I have a feeling it needs cleaning up because it takes almost 1GB on my machine and it seems to contain data belonging to earlier versions of packages. For example, It is further complicated by the fact that /var/pkg/pkg is in the root file system. This means that even if you clean it up from your current boot environment, it will still exist in other boot environments that have been created along the way. It really needs to be in /var/share. http://mgerdts.blogspot.com/2008/03/future-of-opensolaris-boot-environment.html http://defect.opensolaris.org/bz/show_bug.cgi?id=3132 http://defect.opensolaris.org/bz/show_bug.cgi?id=3296 -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Live Upgrade question
On Fri, May 29, 2009 at 3:34 PM, Ivan Ordonez ordonezi...@gmail.com wrote: uname -v shows snv_110. Is that Nevada? Sorry I'm a total noob. What does cat /etc/release say? -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Live Upgrade question
On Fri, May 29, 2009 at 6:40 PM, Ivan Ordonez ordonezi...@gmail.com wrote: OpenSolaris 2009.06 snv_110 X86 Copyright 2009 Sun Microsystems, Inc. All Rights Reserved. Use is subject to license terms. Assembled 19 March 2009 Then you are using OpenSolaris - not SXCE. In the bad-old-days of live upgrade, you had to worry about slices and copying file systems between them. In most systems, this established a practical limit of only a couple boot environments - especially if you wanted to be sure that you always had mirrors. With the good new days of Snap Upgrade (and ZFS enhancements to Live Upgrade as of S10 u6), ZFS comes to the rescue. Creation of a new boot environment is much simpler. You don't have to worry about slices for extra boot environments. With OpenSolaris, it is as simple as... beadm create mynewbe That should take less than a second. The new boot environment is bootable. More details at http://wikis.sun.com/download/attachments/78086473/beadm.1m.txt. But most people will never do that because the packaging system now likes to create boot environments for you. On Monday when OpenSolaris 2009.6 is released it will be at build 112b. You should be able to upgrade your existing machine to it - in a new boot environment - with: Upgrade the ipkg utility to the current version. This changes your active BE $ pfexec pkg install SUNWipkg Create a new BE and upgrade it: $ pfexec pkg image-update On your next reboot you will boot to the newly installed bits. You should also see a grub menu entry for the previous BE in the event that things didn't go well. It's not clear whether you are trying to do this for the advertised reasons for using live upgrade or for the purpose of cloning your system to make another one that looks just like it. If you are trying to clone the system, you probably want to use beadm with the -p option to have the new boot environment be created in a new zpool. That is, it would probably be something like: zpool create newpool /dev/dsk/disk beadm create -p newpool mynewbe (set up grub on disk) I haven't tried this, so there are probably some gotchas when moving to a new system. I'm sure you will write back if/when you stumble across this and discover the fix. ;) -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Transcript from OpenSolaris Livechat on Slashdot...
On Thu, May 28, 2009 at 7:30 PM, Anon Y Mous system5u...@yahoo.com wrote: There was an OpenSolaris live chat with some very eloquent Sun engineers on Slashdot this morning. Does anybody know when and where the transcript will be posted? All opensolaris.com has right now is a link that just goes to slashdot.org on their front page, but no transcript. http://dlc.sun.com/osol/advocacy/opensolaris-slashdot-session-28-05-2009.html FYI: I was the guy who asked if and or when you guys were ever planning on support for Nvidia MCP67 ethernet cards. See following bug reports below: [snip] The advice and discussion in this message makes me think that there are probably others that would like to add follow-on answers or questions. I'm not sure it is practical this time around, but perhaps in the future it would be useful to set up a jive or phpbb forum where each question/answer is the start of a new thread/topic. The posted transcript could then have a link per question to the appropriate thread/topic to allow people to dive a bit deeper than the one sentence question+answer. I would hesitate to blast the osol-discuss list with such traffic. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Solaris 10 10/08 install Crossbow
On Tue, May 5, 2009 at 6:11 AM, James Carlson james.d.carl...@sun.com wrote: Gian Paolo Buono writes: Thanks for your reply, Now I have installed Solaris Express Community Edition and Crossbow work fine ... Can I create a zone with solaris10 on nevada ? I don't believe that's supported yet, but you'd need to contact the Zones community (zones-disc...@opensolaris.org) to get more accurate information. Jerry has just started the discussion on this. http://mail.opensolaris.org/pipermail/zones-discuss/2009-April/004897.html -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Question about (Open)Solaris VM - why is swap required?
On Sat, Apr 25, 2009 at 2:01 AM, Shawn Walker swal...@opensolaris.org wrote: Anon Y Mous wrote: as we tested scaling and put more load on the servers (i.e. allowing Apache to spawn more children), we were surprised to see that the system doesn't tend to go below 8Gb (half) of the available RAM. Further requests for memory allocation failed (can't fork new processes including new sshd, vmstat, top and bash). This may also be helpful: http://blogs.sun.com/jimlaurent/entry/solaris_faq_myths_and_facts The first comment in this blog entry probably reflects the behavior that you are seeing. In comparison, Linux will allow overallocation of memory. If things try to touch more memory+swap space than the system has, Linux will invoke its out of memory killer. See, for example, http://lwn.net/Articles/104179/. This behavior can be turned off on Linux, but I believe that it is enabled by default on most distros. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Can mysql be slower in a zone?
On Fri, Apr 10, 2009 at 9:19 AM, Mike DeMarco mikej...@yahoo.com wrote: You should only see a very little performace hit when using zones. Maybe 1 or 2 %. Are you using exclusive NIC to move the TCP stack down into the zone? What evidence do you have for that? There should be no overhead for a zone. If you have workloads that show a performance hit in zones, please bring it up at zones-discuss. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] How to move the os to anther disk (no mirroring)
On Fri, Mar 20, 2009 at 8:04 PM, Harry Putnam rea...@newsguy.com wrote: I want to move my OS (rpool) to a different disk. It's all a pretty new install but I have gotten enough updates and config edits in place that I'd not like to re-install from 2008.11 101b. Can I use `dd' for that task or is that likely to have problems with zfs? Is there a better way? Probably so, using zfs to do the copying of only the used blocks. It goes something like: 1. Physically install the new disk. 2. Partition it as you would like (assuming c0t1s0 is the new destination slice) 3. zpool attach -f rpool c0t0s0 c0t1s0 4. installgrub -m /boot/grub/stage1 /boot/grub/stage2 /dev/rdsk/c0d1s0 5. zpool detach rpool c0t0s0 I used http://darkstar-solaris.blogspot.com/2008/09/zfs-root-mirror.html as for hints on the process. There may be one more magical step (possibly a reboot?) to get it to recognize that rpool is now 500 GB instead of 25 GB. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] No login screen after creating .xorg.conf
On Mon, Mar 2, 2009 at 4:33 PM, Marcelo M. Garcia marcelo.maia.gar...@googlemail.com wrote: Hi After updating my system (Dell XPS M1530 with nvidia gt8600m) from 101 to 108, the keyboard layout stopped to work. This is a known bug. The work around was to create a /etc/X11/xorg.conf and insert the keyboard configuration. I created my /etc/X11/xorg.conf using /usr/bin/nvidia-xconfig. Then edited the file like in [1], but now when the computer starts, all I get is the splash screen with a sliding bar. It never reaches the login screen. I tried to boot in text mode editing the boot line[2] to remove the console=graphics, but it only made the matters worst. I found in virtualbox that removing console=graphics and adding -s (single user) allowed me to boot. I then disabled gdm (svcadm disable -t gdm) and ran svcadm milestone all (or something like that). Upon starting X from the command line I found that it was complaining about an rgb line in xorg.conf. Removing the offending rgb line made things all better. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] MAX_ZONEID in usr/src/uts/common/sys/zone.h
On Wed, Dec 10, 2008 at 3:51 PM, Dennis Clarke [EMAIL PROTECTED] wrote: Any particular reason why we have #define MAX_ZONEID in zone.h ? I am searching for the definition for zoneid_t but I am guessing that it is bigger than an unsigned 16 bit integer. Given sufficient resources I would guess that a machine could have greater than zones defined and configured regardless if they are running or not. zone id's are like PID's - each zone may have a different zone id each time it is booted. It seems as though this value is used to initialize a id_space, which is described at: http://src.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/uts/common/os/id_space.c#33 -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] IMPT: Do not use SXCE Build 102
On Sat, Nov 15, 2008 at 10:49 PM, John Beck [EMAIL PROTECTED] wrote: I would recommend the tip. If that wouldn't work for you, then please explain what further criteria you might have that I can't guess. :-) Thanks. I just wanted to be sure that tip had the fix (or unbreak) and I didn't need to go back to before the bug was introduced. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] IMPT: Do not use SXCE Build 102
On Sat, Nov 15, 2008 at 1:14 PM, Derek Cicero [EMAIL PROTECTED] wrote: Due to the following bug, I have removed build 102 from the Download page. 6771840 zpool online on ZFS root can panic system It apparently may cause data corruption and may have been implicated in damage to one or more systems that have upgraded to build 102 or beyond. Is this related to: http://hg.genunix.org/onnv-gate.hg/rev/ef46e52dcf97 backout 6752810: causes data corruption on zfs root pool I'm trying to figure out what is a reasonably safe changeset to use as I am building ON. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] How to know free memory of solaris system?
On Fri, Oct 31, 2008 at 9:30 PM, zhihui Chen [EMAIL PROTECTED] wrote: intel7# vmstat 2 kthr memorypagedisk faults cpu r b w swap free re mf pi po fr de sr rm s0 s1 -- in sy cs us sy id 5 0 0 7588092 5866084 8 163 7 6 18 0 124 -20625 12 -0 0 3205 194187 50279 10 3 87 0 0 0 7935004 6799272 7 45 0 0 0 0 0 0 9 0 0 431 228 390 0 0 100 0 0 0 7934852 6799140 0 5 0 0 0 0 0 0 0 0 0 417 236 341 0 0 100 intel7# java -d64 -Xms8g -Xmx8g -version Error occurred during initialization of VM Could not reserve enough space for object heap Memory allocations place a reservation against swap. If vmstat's swap column says you have less than 8 GB of swap available, you will not be able to reserve 8 GB of memory. Note that even if the system shows 1 TB in the free column, memory allocations will fail if the swap column is smaller than what you are trying to allocate. The swap column shows the amount of RAM + swap device space that has not already been reserved. I'm not sure why the ::memstat and vmstat are in disagreement on the size of the free list. My understanding is that they should be the same. According to vmstat(1M): memoryReport on usage of virtual and real memory. swapavailable swap space (Kbytes) freesize of the free list (Kbytes) Perhaps the ZFS ARC has changed the usefulness of the ::memstat output. You may get a more focused audience on perf-discuss. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] migrators question: virtual terminals?
On Tue, Sep 30, 2008 at 9:59 AM, Ron Halstead [EMAIL PROTECTED] wrote: Not now, but: http://opensolaris.org/os/community/on/flag-days/pages/2008092402/; The xorg.tar.gz, when gunzipped and untared results in tar: directory checksum error gnu tar (/usr/sfw/bin/gtar) and Sun's tar (/usr/bin/tar) use different mechanisms for encoding long path names. If one gives this error, try the other. If they both fail, then the archive is probably really corrupt. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] ACER Aspire One and OpenSolaris B99, fdisk
On Sun, Sep 28, 2008 at 1:48 PM, andrew [EMAIL PROTECTED] wrote: Acer is shipping laptops with no recovery solution as standard? That sounds well dodgy to me. Andrew. In the last couple of weeks I've bought two cheap ($399) laptops (Compaq, Acer). Neither came with a recovery DVD, but both came with a program to create a recovery DVD. I haven't tested recovery with that DVD for either. I found a GNU Parted live CD image (google helped) and quite painlessly shrank the NTFS partition on the Compaq to make room for an OpenSolaris partition. It installed just fine but I didn't get wifi to work. That machine went back due to problems with the power connector. Huge kudos go out to Office Depot. They have a 14 day no-restocking fee return policy and gave me no grief for returning a box on day 14. They only asked whether it was defective so they knew whether it needed to go back to HP or not. The Acer (Extensa 4420) came with a 10 Gig hidden partition, a 50 GB C drive, and a 50 GB D drive. It seems as though everything important to windows is in the C drive and the D drive is the default location for Acer's backup utility for backups. Once I get a chance the D drive will go to better use. My first boot of the build 95 snapshot of OpenSolaris says that I will need to do battle with networking (wifi and wired) as well as sound. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] FSS - What % do you need to give the global zone.
On Tue, Aug 26, 2008 at 2:59 PM, Robert Milkowski [EMAIL PROTECTED] wrote: Generally try to select numbers in such a way for all local zones that all shares add up to 100 - that way your number of shares is your percentage of all CPUs you want to guarantee for that zone. My approach is a bit different. Since various servers I support have different amounts of oomph, I have invented a term called Zone Power Unit (zpu). I started out with a particular configuration of a V240 that was popular at the time and declared that it had 10 zpus. I figured I would never put more than 10 zones on such a server. As time has gone on and there are other more powerful servers, I have assigned various zpu values for them. Consider moving a zone that used half of a V240 to a T2000. On the V240 it was allocated half of the server. On the T2000 it is allocated quite a bit less than half while being allocated 5 zpus on each. My zpu values are based upon mvalues, received from Sun under NDA so I can't share the actual values here. If I didn't have the mvalues from Sun, I'd probably go looking for results at spec.org, sap.com, etc. to understand the relative performance of the various systems. No matter what it won't be perfect because we are relying on benchmarks. However, it does give a good way to understand the approximate utilization and capacity of various servers on a normalized basis. As for the initial question - I would agree that allocating a lot of shares to the global zone (that runs no applications) is a good safeguard to ensure that you can do administration. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] which package provides math.h?
On Wed, Aug 20, 2008 at 6:24 PM, Matt Harrison [EMAIL PROTECTED] wrote: I probably should have mentioned, I'm running SXCE (snv_95). I've tried all the TERM=xtermc, TERM=xterm-color stuff, and I'm still in black and white. I have the following in my .vimrc: :if has(terminfo) : set t_Co=8 : set t_Sf=ESC[3%p1%dm : set t_Sb=ESC[4%p1%dm : syntax on :else : set t_Co=8 : set t_Sf=ESC[3%dm : set t_Sb=ESC[4%dm : syntax on :endif Replace ESC with an escape character (^Vesc). -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] installing OpenSolaris not easy-how does someone
On Fri, Aug 15, 2008 at 6:57 AM, Marc [EMAIL PROTECTED] wrote: Hi, Trying to find that but meanwhile is this correct? copy/b osol-0811-95.iso + md5sums md5sums_95.txt osol-0811-95a.iso or copy/b osol-0811-95.iso + md5sums_95.txt osol-0811-95a.iso I assume you are asking about the DOS (Windows CMD prompt) copy command. You do not need to do this and if you do this you will end up with a corrupt iso image named osol-0811-95a.iso. The md5sums.txt file is used only if you want to verify that you downloaded a good copy of the .iso file before burning it to a CD. The way that I would do it is with: openssl md5 osol-0811-95.iso Then I would compare the md5 sum (long string of mostly letters and digits that looks like gibberish) displayed by openssl with the contents of md5sums_95.txt. In practice, I only do this when I suspect that there is a problem with the file that I need to verify. In other words, I almost never do this. I do not know of a command that comes with Windows that will compute an md5sum. You should skip this test and instead simply burn osol-0811-95.iso to a CD. Burn it as a pre-created image. If after you are done burning the CD, you insert the CD into a windows box and it tells you that there is one file on the CD named osol-0811-95.iso, you did it wrong. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] blastwave.org - any info?
On Thu, Aug 7, 2008 at 6:34 AM, Dennis Clarke [EMAIL PROTECTED] wrote: You can not restore and then distribute all those binary packages unless you also have all the sources and source changes as well as the rights to infringe on both copyright and license. Which you don't. Simply be careful of what you choose to step into. Just a hunch, but the Coder's Rights Project or other EFF efforts may be of interest for legal assistance. http://www.eff.org/issues/coders -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Reading ext3 filesystem from redhat branded zone?
On Sat, Aug 2, 2008 at 4:55 AM, Johan Hartzenberg [EMAIL PROTECTED] wrote: On Sat, Aug 2, 2008 at 2:30 AM, Krenz von Leiberman [EMAIL PROTECTED] wrote: Is it possible to read an external hard drive with an ext3 filesystem from a redhat branded zone? Thanks in advance. No. Branded zones run Solaris, but with Linux-lookalike APIs. Therefore, no Linux kernel, and subsequently no Linux (device) drivers are present. But with FUSE, it seems as though the ext3 file system could be mounted from the Solaris global zone and at least lofs mounted into the Linux branded zone. http://blogs.sun.com/pradhap/entry/mount_ntfs_ext2_ext3_in Google solaris fuse ext3 for some other possibly useful pages. I haven't tried tried this myself... -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Reading ext3 filesystem from redhat branded zone?
On Sat, Aug 2, 2008 at 7:21 AM, Mike Gerdts [EMAIL PROTECTED] wrote: On Sat, Aug 2, 2008 at 4:55 AM, Johan Hartzenberg [EMAIL PROTECTED] wrote: On Sat, Aug 2, 2008 at 2:30 AM, Krenz von Leiberman [EMAIL PROTECTED] wrote: Is it possible to read an external hard drive with an ext3 filesystem from a redhat branded zone? Thanks in advance. No. Branded zones run Solaris, but with Linux-lookalike APIs. Therefore, no Linux kernel, and subsequently no Linux (device) drivers are present. But with FUSE, it seems as though the ext3 file system could be mounted from the Solaris global zone and at least lofs mounted into the Linux branded zone. http://blogs.sun.com/pradhap/entry/mount_ntfs_ext2_ext3_in Google solaris fuse ext3 for some other possibly useful pages. I haven't tried tried this myself... I hit send too early... There is also this project that seems to have made some headway: http://opensolaris.org/os/project/ext3/ http://opensolaris.org/os/project/ext3/Documentation/quick_start/ -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Can Solaris be discussed here?
On Sat, Aug 2, 2008 at 4:23 PM, Ming Kin Lai [EMAIL PROTECTED] wrote: If asking for a way to report a bug resembles looking for free support, well, I don't know what to say. ?If I have a product and care about it, I would love people giving feedback and tell me what's wrong about it. ?I too have to eat, pay rent and gas, etc. ?And I basically would provide free service to the company by reporting a bug to it. I would suggest trying to reproduce the problem in OpenSolaris, Solaris Express, etc. If it still exists and you can't find the bug by searching at bugs.opensolaris.org, file it there. If the bug has been fixed in OpenSolaris, there is a pretty good chance that Sun knows about it and has not chosen to (or chosen not to) backport it to Solaris 10. If the bug still exists in OpenSolaris but you can't find a bug report, you can file one at bugs.opensolaris.org. Since the code bases are closely related Sun has a pretty good idea that it affects more than just OpenSolaris. I too am often annoyed that my pet bug doesn't get fixed in the release I need to run in production. However, I have embraced involvement (honest feedback, evaluations of proposals, code contributions) in OpenSolaris as my means to limiting the time that bugs bother me. It doesn't mean that I always get my way, but when approached tactfully it almost always means that I am heard by those that are best able to cause changes to happen. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] [osol-bugs] man: segmentation fault (snv_91)
On Sat, Jun 21, 2008 at 2:24 PM, Mike Gerdts [EMAIL PROTECTED] wrote: On Fri, Jun 20, 2008 at 11:56 AM, Mike Gerdts [EMAIL PROTECTED] wrote: I've sent mail to request-sponsor to get someone to help me with the putback of the fix. I expect that I will have a fix ready for code review by Monday. The fix is ready for code review at http://cr.opensolaris.org/~mgerdts/6717067/. I'll send a separate request for code review to on-discuss as well. The fix was committed on July 23 and should appear in build 96. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] SVM thru jumpstart
On Mon, Jul 28, 2008 at 4:36 PM, AV [EMAIL PROTECTED] wrote: Based on the above 'profile' setup, jumpstart gives me below numbering for the subdisks. These commands would be auto-populated in /tmp/create_mirror. - metainit d0 -m d5 d6 metainit d1 -m d7 d8 metainit d2 -m d9 d10 metainit d3 -m d11 d12 metainit d4 -m d13 d14 -- This looks like the way that Solaris 9 initially did it. Solaris 10 and newer (and possibly even later S9 releases) create metadevice names the way you want. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Where can I find a document describing OpenSolaris commands?
On Fri, Jul 25, 2008 at 2:30 AM, Thommy M. [EMAIL PROTECTED] wrote: Varuna Seneviratna wrote: Where can I find a document describing OpenSolaris commands? Do as the rest of us did, start with the man command man man Unfortunately, opensolaris is missing many man pages. Those missing pages can be found at http://docs.sun.com/app/docs/coll/40.17. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Any lsof tool for nv ?
On Tue, Jul 15, 2008 at 8:05 AM, Richard L. Hamilton [EMAIL PROTECTED] wrote: # List open files with link count 1 (which could account # for du != df). This is just part of it, the whole list is # too long. Also, there aren't any really large ones just # now, so it's not an actual concern at the moment. And of # course, Jive messed up the columns; in reality, they do line up. Without lsof, you can do this with: # find /proc/*/fd -type f -links 0 -ls If you want to look inside the file to get an idea of what it is: # less /proc/$pid/fd/$fd If you want to truncate that 1 GB log file that has been deleted: # cp /dev/null /proc/$pid/fd/$fd This is not meant to lessen the importance of having lsof (really, lsof not some almost work-a-like) in OpenSolaris. It is a very helpful tool and is a common part of the known tool set of third party support organizations. That is, having this tool makes it so that ISV and freeware software is more easily supportable on OpenSolaris. I understand that lsof does bad things to get the data that it gets. It would likely stop doing that just about the time that stable interfaces are provided to make it unnecessary to go prodding around in /dev/kmem. I guess since I am complaining, I should add this to the list of projects for me to look into in the future... -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Any lsof tool for nv ?
On Tue, Jul 15, 2008 at 10:12 AM, Shawn Walker [EMAIL PROTECTED] wrote: The main problem with lsof is that it will break often as kernel structures, etc. change. Which is why I suggested a stable interface is needed to mask these changes. lsof needs a very small amount of the data that is in the kernel data structures. The data that it needs (path, ip address, port, shared memory segment id, etc.) are things that are quite unlikely to go away over time. The data structure that the kernel uses to keep track of them may change. It might be easiest to find a wrapper script that emulates lsof to put around pfiles. From proc(1), the page that comes up in response to man pfiles. ATTRIBUTES ... Human Readable Output is Unstable. Options are Evolving. Such a wrapper script will have the same problems as lsof (perhaps with less frequency) while not having feature parity. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] net prob
On Sun, Jul 13, 2008 at 6:52 AM, john kroll [EMAIL PROTECTED] wrote: Thanks for your reply but now E521: /usr/ccs/bin/makenot working ?? You need to provide more details of what you are trying to do, the exact steps you are taking, and the exact error messages you are seeing. So far, I have seen nothing that would make me think that you are having any sort of net prob (network problem?). -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] net prob
On Sat, Jul 12, 2008 at 9:47 PM, john kroll [EMAIL PROTECTED] wrote: sxce93 #`isainfo -n` amd64:cannot execute? Get rid of the backticks # isainfo -n OR # echo isainfo -n reports: `isainfo -n` -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] [pkg-discuss] samba 3.2.0 on os200805 won't get past configure
On Fri, Jul 11, 2008 at 7:55 AM, Dennis Clarke [EMAIL PROTECTED] wrote: On Fri, Jul 11, 2008 at 3:24 AM, Moinak Ghosh [EMAIL PROTECTED] wrote: On Fri, Jul 11, 2008 at 3:07 AM, Dennis Clarke [EMAIL PROTECTED] wrote: I'm trying to figure out .. what is wrong. It could be me or it could be my system state .. not too sure. I cna build samba 3.2.0 on a trusty old Solaris 8 x86 machine without much fanfare : [...] but with os200805 I can not get past configure. There are complaints about libpam at the very least : http://www.blastwave.org/dclarke/stuff/samba_on_os200805.txt any thoughts ? The contents of config.log will be more helpful in figuring this. it is somewhat long http://www.blastwave.org/dclarke/stuff/samba-3.2.0-build_config_num.log with line numbers added ... I'll start surfing along it to see the issue. It is really a problem with kerberos (/bin/krb5-config), not PAM. 153029 configure:63539: checking for pam_get_data in -lpam 153030 configure:63572: cc -o conftest -I/usr/include/kerberosv5 -O -D_SAMBA_BUILD_=3 -I/usr/include/kerberosv5 -I/opt/csw/include -D_LARGEFILE_SOURCE -D_REENTRANT -D_FILE_OFFSET_BITS=64 -Iinclude -I../samba-3.2.0/source/include -I. -I../samba-3.2.0/source -I../samba-3.2.0/source/lib/replace -I../samba-3.2.0/source/lib/talloc -I../samba-3.2.0/source/lib/tdb/include -I../samba-3.2.0/source/libaddns -I../samba-3.2.0/source/librpc -I../samba-3.2.0/source/popt -I/opt/csw/include -DLDAP_DEPRECATED /bin/krb5-config: Unknown option `gssapi' -- use `--help' for usage -lthread -L./bin -L/opt/csw/lib conftest.c -lpam -lresolv -lnsl -lsocket -liconv 5 153031 conftest.c, line 518: warning: statement not reached 153032 ld: fatal: file /bin/krb5-config:: open failed: No such file or directory 153033 ld: fatal: file Unknown: open failed: No such file or directory 153034 ld: fatal: file option: open failed: No such file or directory 153035 ld: fatal: file `gssapi': open failed: No such file or directory 153036 ld: fatal: file use: open failed: No such file or directory 153037 ld: fatal: file `--help': open failed: No such file or directory 153038 ld: fatal: file for: open failed: No such file or directory 153039 ld: fatal: file usage: open failed: No such file or directory 153040 ld: fatal: File processing errors. No output written to conftest 153041 configure:63578: $? = 1 153042 configure: failed program was: Grumble. krb5-config writes errors to stdout. $ /bin/krb5-config gssapi /bin/krb5-config: Unknown option `gssapi' -- use `--help' for usage $ /bin/krb5-config gssapi /dev/null Looks to me like line 92 of krb5-config needs 12 appended. $ cat -n /usr/bin/krb5-config | grep for.usage 92 echo $0: Unknown option \`$1' -- use \`--help' for usage -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] [pkg-discuss] samba 3.2.0 on os200805 won't get past configure
On Thu, Jul 10, 2008 at 4:37 PM, Dennis Clarke [EMAIL PROTECTED] wrote: but with os200805 I can not get past configure. There are complaints about libpam at the very least : http://www.blastwave.org/dclarke/stuff/samba_on_os200805.txt any thoughts ? FWIW, things seem to progress OK for me. At one time I did pkg install ss-dev which was supposed to bring in the studio compilers and other development goodness. ./configure --prefix=/tmp --with-pam things seem to go OK. ... checking whether to try PAM support... yes checking security/pam_appl.h usability... yes checking security/pam_appl.h presence... yes checking for security/pam_appl.h... yes checking pam/pam_appl.h usability... no checking pam/pam_appl.h presence... no checking for pam/pam_appl.h... no checking for pam_get_data in -lpam... yes checking for security/pam_modules.h... yes checking for pam/pam_modules.h... no checking security/pam_ext.h usability... no checking security/pam_ext.h presence... no checking for security/pam_ext.h... no checking security/_pam_macros.h usability... no checking security/_pam_macros.h presence... no checking for security/_pam_macros.h... no checking pam/pam_ext.h usability... no checking pam/pam_ext.h presence... no checking for pam/pam_ext.h... no checking pam/_pam_macros.h usability... no checking pam/_pam_macros.h presence... no checking for pam/_pam_macros.h... no checking for pam_vsyslog... no checking whether to use PAM support... yes checking whether to have PAM MODULES support... yes checking whether to use pam_smbpass... no checking for library containing crypt... none required checking whether to use NISPLUS_HOME... no ... -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Question on how Swap works
On Mon, Jul 7, 2008 at 6:15 AM, Frank Mancini [EMAIL PROTECTED] wrote: Thanks for your reply. So are you saying the reason why I'm using so much swap is because one of my applications that I'm starting up is reserving this space? Yes. Take a look at the output of prstat -s size. It will show you how much memory is reserved (SIZE) and the amount resident in RAM (RSS). I guess I always thought it reserved regular physical memory first, then swapped later when physical runs out. ?So realistically, I can run out of swap without running out of physical memory first? I've seen it happen. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Question on how Swap works
On Mon, Jul 7, 2008 at 7:01 AM, Lars Tunkrans [EMAIL PROTECTED] wrote: One more usage of swap is TMPFS . If you have lots of files in /tmp They use the swap area . //Lars Remember swap is RAM + swap devices. The notion that /tmp is mounted on swap says that the backing store for that file system may be RAM and/or swap devices. For performance reasons, RAM is preferred. It is only if you are running short on RAM that swap devices should be seeing any I/O operations. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Question on how Swap works
On Fri, Jul 4, 2008 at 11:18 AM, Frank Mancini [EMAIL PROTECTED] wrote: Quick question. On my server, I have: 16 gigs of real memory 5 gigs are free 32 gigs of swap 28 gigs of swap used Doesn't memory fill the real physical first and how come I'm using so much swap? When an application asks for memory a swap reservation is performed. This reservation is immediately seen in the swap column of vmstat. In this context, swap means the combined space of RAM and swap devices. Since none of the memory has been touched yet, there is no reason to reserve RAM for it - in fact there is a pretty good chance that some of it will never be touched before the process exits. An example of where I have seen this is when some application manual says to set the java heap size (-xMx) to X GB, but the application running in the JVM never touches much of that. As soon as java starts, a bit more than X GB of memory is reserved but a small fraction of it is touched - free swap is decreased but free RAM remains mostly constant. As the application uses memory that is reserved the free swap remains constant and the free RAM decreases. -- Mike Gerdts http://mgerdts.blogspot.com/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
