[osol-discuss] lpr dumps core when called from firefox in snv_160
When printing from firefox using lpr in snv_160, lpr dumps core: datsunx$ adb /usr/lib/lp/bin/lpr core core file = core -- program ``/usr/lib/lp/bin/lpr'' on platform i86pc SIGSEGV: Segmentation Fault $c libpapi-common.so.0`regvalue+0x16(fedb5ad4, fedb5ad8, 0, 4, 806b4a0, fef6e000) libpapi-common.so.0`_process_value+0xed(fedb5ad4, 8046eac, fedb1970, fedb3b11) libpapi-common.so.0`_papiAttributeFromStrings+0xcb(80669d8, 1, 806a6c8, 806ab28 ) libpapi-common.so.0`_parse_attribute_list+0x1fc(80669d8, 1, 8046fb0, 8065eb0) libpapi-common.so.0`papiAttributeListFromString+0x2c(80669d8, 1, 8046fb0, 100) psm-lpd.so.1`papiJobStreamOpen+0x217(8066200, 806a158, 80668e0, 0, 804714c, fedd58a8) libpapi.so.0`papiJobStreamOpen+0x6e(8066e50, 8047e89, 80668e0, 0, 804714c, 0) jobSubmitSTDIN+0x30(8066e50, 8047e89, 80475ac, 3, 80668e0, 80475a4) main+0x546(1, 80475f4, 80475fc, 8051e9f) _start+0x7d(1, 8047778, 0, 804778c, 804779d, 80477ab) datsunx$ dbx /usr/lib/lp/bin/lpr core For information about new features see `help changes' To remove this message, put `dbxenv suppress_startup_message 7.5' in your .dbxrc Reading lpr core file header read successfully Reading ld.so.1 Reading libc.so.1 Reading en_US.UTF-8.so.3 Reading methods_unicode.so.3 Reading libpapi.so.0 Reading libpapi-common.so.0 Reading nss_user.so.1 Reading nss_files.so.1 Reading psm-lpd.so.1 program terminated by signal SEGV (access to address exceeded protections) 0xfedb37be: regvalue+0x0016:movb $0x,(%eax,%edx) (dbx) where =[1] regvalue(0xfedb5ad4, 0xfedb5ad8, 0x0, 0x4, 0x806b4a0, 0xfef6e000), at 0xfedb37be [2] _process_value(0xfedb5ad4, 0x8046eac, 0xfedb1970, 0xfedb3b11), at 0xfedb38b1 [3] _papiAttributeFromStrings(0x80669d8, 0x1, 0x806a6c8, 0x806ab28), at 0xfedb3bcb [4] _parse_attribute_list(0x80669d8, 0x1, 0x8046fb0, 0x8065eb0), at 0xfedb3e24 [5] papiAttributeListFromString(0x80669d8, 0x1, 0x8046fb0, 0x100), at 0xfedb3e9c [6] papiJobStreamOpen(0x8066200, 0x806a158, 0x80668e0, 0x0, 0x804714c, 0xfedd58a8, 0x1, 0xfedd2b79), at 0xfed03223 [7] papiJobStreamOpen(0x8066e50, 0x8047e89, 0x80668e0, 0x0, 0x804714c, 0x0, 0x0, 0xfefe45c9), at 0xfedd2bd6 [8] jobSubmitSTDIN(0x8066e50, 0x8047e89, 0x80475ac, 0x3, 0x80668e0, 0x80475a4, 0x0, 0x8053998), at 0x80532bc [9] main(0x1, 0x80475f4, 0x80475fc, 0x8051e9f), at 0x805252a (dbx) quit Is this a known bug? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] smbd fails because libavahi-core.so.6 is missing
I upgraded to snv_160 and the Samba daemon (smbd) fails to start because it depends on libavahi-core.so.6, which is missing. [ Mar 5 23:16:15 Executing start method (/usr/sbin/smbd -D). ] ld.so.1: smbd: fatal: libavahi-core.so.6: open failed: No such file or directory Is this a known bug? Did I do something wrong? The avahi package only seems to include libavahi-core.so.7. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] fork1() fails with ENOMEM
I've made a little progress on this problem. Using the attached dtrace script, I get output that looks like this: dtrace: script 'fork.d' matched 144 probes CPU FUNCTION 0 - as_dup 0- hat_dup 0- hat_dup 0 0- segvn_dup 0- segvn_dup 0 0- segvn_dup 0 - anon_resvmem size 4096 takemem 1 tryhard 1 0 - anon_resvmem 1 0 - anon_create 0 - anon_create-997832123064 0 - anon_dup 0 - anon_dup 0 0- segvn_dup 0 ... 0- segvn_dup 0 - anon_resvmem size 119177216 takemem 1 tryhard 1 0 - anon_resvmem 1 0 - anon_create 0 - anon_create-997820303808 0- segvn_dup 0 0- segvn_dup 0 - anon_resvmem size 113246208 takemem 1 tryhard 1 0 - anon_resvmem 0 0- segvn_dup 12 0 - as_dup 12 0 = forksys FORK errno 12 There's 85 calls to segvn_dup, all of which succeed except the last one above. I tried setting anon_debug using mdb, but I'm not even sure if the debug code is compiled into the kernel. FYI, swap -s says: # swap -s total: 2816356k bytes allocated + 496608k reserved = 3312964k used, 730976k available There are no messages in /var/adm/messages about running out of swap space. The total of all the sizes for the anon_resvmem calls is over 800MB. If I'm just plain running out of swap space, shouldn't there be a message in /var/adm/messages? And that doesn't explain why this program fails in the same way on another machine with 150GB of swap space. Is there an easy way to add more swap space to a system that's using zfs for swap space? I tried mkfile and swap -a but that doesn't work. Any other ideas for how to debug this? #!/usr/sbin/dtrace -Fs #pragma D option cleanrate=1000hz #pragma D option nspec=4 syscall::fork*:entry { self-s = speculation(); speculate(self-s); } syscall::fork*:return /self-s/ { speculate(self-s); printf(FORK errno %d\n, errno); } syscall::fork*:return /self-s arg1 == -1 errno != 0/ /* fork1 fails! */ { commit(self-s); self-s = 0; } syscall::fork*:return /self-s !(arg1 == -1 errno != 0)/ { discard(self-s); self-s = 0; } fbt::*dup:entry /self-s/ { speculate(self-s); } fbt::*dup:return /self-s/ { speculate(self-s); trace(arg1); } fbt::anon_resvmem:entry /self-s/ { speculate(self-s); printf(size %d takemem %d tryhard %d\n, arg0, arg1, arg3); } fbt::anon_resvmem:return /self-s/ { speculate(self-s); trace(arg1); } fbt::anon_create:entry /self-s/ { speculate(self-s); } fbt::anon_create:return /self-s/ { speculate(self-s); trace(arg1); } fbt::anon_private:entry /self-s/ { speculate(self-s); } fbt::anon_private:return /self-s/ { speculate(self-s); trace(arg1); } ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] fork1() fails with ENOMEM
This isn't really specific to OpenSolaris since it also happens on Solaris 10, but maybe someone here can give me some ideas? I have a java program that is failing because it does something that calls fork1(), which fails with ENOMEM (confirmed using truss): 27010/2:fork1() Err#12 ENOMEM I've used both the 32-bit and 64-bit versions of the JVM. Using the 64-bit version, I see the heap expand to just over 4GB before it forks: 27010/30: brk(0x107DE3D90)= 0 I have almost 160 GB of swap space: $ swap -s total: 806336k bytes allocated + 167872k reserved = 974208k used, 159746184k available It doesn't seem like it can possibly be running out of swap space. What other reasons would cause fork1() to fail with ENOMEM? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] fork1() fails with ENOMEM
Bart Smaalders wrote on 02/ 4/11 11:49 AM: On 02/04/11 11:33, Bill Shannon wrote: This isn't really specific to OpenSolaris since it also happens on Solaris 10, but maybe someone here can give me some ideas? I have a java program that is failing because it does something that calls fork1(), which fails with ENOMEM (confirmed using truss): 27010/2: fork1() Err#12 ENOMEM I've used both the 32-bit and 64-bit versions of the JVM. Using the 64-bit version, I see the heap expand to just over 4GB before it forks: 27010/30: brk(0x107DE3D90) = 0 I have almost 160 GB of swap space: $ swap -s total: 806336k bytes allocated + 167872k reserved = 974208k used, 159746184k available It doesn't seem like it can possibly be running out of swap space. What other reasons would cause fork1() to fail with ENOMEM? ___ The vm system is rather fond of using ENOMEM as a generic error bucket. If you have a way of reproducing the problem, a bit of DTrace will quickly turn up the reason. I assume it's only the 32 bit version that is failing to fork, right? Sorry, I wasn't clear. The 64-bit version fails as well. With the 32-bit version I limit the Java heap to 2GB. It doesn't run out of heap, but the fork fails. I'm sure dtrace will solve every problem I have, but I don't know how to use it to solve this problem! :-( Hints? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] fork1() fails with ENOMEM
Edward Martinez wrote on 02/ 4/11 01:52 PM: On 02/04/11 13:06, Bill Shannon wrote: Bart Smaalders wrote on 02/ 4/11 11:49 AM: On 02/04/11 11:33, Bill Shannon wrote: This isn't really specific to OpenSolaris since it also happens on Solaris 10, but maybe someone here can give me some ideas? I have a java program that is failing because it does something that calls fork1(), which fails with ENOMEM (confirmed using truss): 27010/2: fork1() Err#12 ENOMEM I've used both the 32-bit and 64-bit versions of the JVM. Using the 64-bit version, I see the heap expand to just over 4GB before it forks: 27010/30: brk(0x107DE3D90) = 0 I have almost 160 GB of swap space: $ swap -s total: 806336k bytes allocated + 167872k reserved = 974208k used, 159746184k available It doesn't seem like it can possibly be running out of swap space. What other reasons would cause fork1() to fail with ENOMEM? ___ The vm system is rather fond of using ENOMEM as a generic error bucket. If you have a way of reproducing the problem, a bit of DTrace will quickly turn up the reason. I assume it's only the 32 bit version that is failing to fork, right? Sorry, I wasn't clear. The 64-bit version fails as well. With the 32-bit version I limit the Java heap to 2GB. It doesn't run out of heap, but the fork fails. I'm sure dtrace will solve every problem I have, but I don't know how to use it to solve this problem! :-( Hints? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org Hi, Please excuse me for getting into the discussion. I think in solaris 10 the /tmp file is place by default in swap, if /tmp is full it may give that error. I found an article on that theory. http://nilesh-joshi.blogspot.com/2010/03/tmp-file-system-full-swap-space-limit.html /tmp isn't full: $ df -h /tmp Filesystem size used avail capacity Mounted on swap 152G 384K 152G 1%/tmp It's hard to tell if it runs out of space during the execution of the process but I doubt it. I don't get any console messages saying the filesystem is full. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] fork1() fails with ENOMEM
Edward Martinez wrote on 02/ 4/11 04:11 PM: On 02/04/11 14:07, Bill Shannon wrote: Edward Martinez wrote on 02/ 4/11 01:52 PM: On 02/04/11 13:06, Bill Shannon wrote: Bart Smaalders wrote on 02/ 4/11 11:49 AM: On 02/04/11 11:33, Bill Shannon wrote: This isn't really specific to OpenSolaris since it also happens on Solaris 10, but maybe someone here can give me some ideas? I have a java program that is failing because it does something that calls fork1(), which fails with ENOMEM (confirmed using truss): 27010/2: fork1() Err#12 ENOMEM I've used both the 32-bit and 64-bit versions of the JVM. Using the 64-bit version, I see the heap expand to just over 4GB before it forks: 27010/30: brk(0x107DE3D90) = 0 I have almost 160 GB of swap space: $ swap -s total: 806336k bytes allocated + 167872k reserved = 974208k used, 159746184k available It doesn't seem like it can possibly be running out of swap space. What other reasons would cause fork1() to fail with ENOMEM? ___ The vm system is rather fond of using ENOMEM as a generic error bucket. If you have a way of reproducing the problem, a bit of DTrace will quickly turn up the reason. I assume it's only the 32 bit version that is failing to fork, right? Sorry, I wasn't clear. The 64-bit version fails as well. With the 32-bit version I limit the Java heap to 2GB. It doesn't run out of heap, but the fork fails. I'm sure dtrace will solve every problem I have, but I don't know how to use it to solve this problem! :-( Hints? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org Hi, Please excuse me for getting into the discussion. I think in solaris 10 the /tmp file is place by default in swap, if /tmp is full it may give that error. I found an article on that theory. http://nilesh-joshi.blogspot.com/2010/03/tmp-file-system-full-swap-space-limit.html /tmp isn't full: $ df -h /tmp Filesystem size used avail capacity Mounted on swap 152G 384K 152G 1% /tmp It's hard to tell if it runs out of space during the execution of the process but I doubt it. I don't get any console messages saying the filesystem is full. Hi, Interesting... it may be running out of heap space, i think limits can be checked with ulimit -a $ ulimit -a time(seconds)unlimited file(blocks) unlimited data(kbytes) unlimited stack(kbytes)8192 coredump(blocks) unlimited nofiles(descriptors) 256 vmemory(kbytes) unlimited ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] fork1() fails with ENOMEM
Edward Martinez wrote on 02/04/2011 07:22 PM: On 02/04/11 16:19, Bill Shannon wrote: Edward Martinez wrote on 02/ 4/11 04:11 PM: On 02/04/11 14:07, Bill Shannon wrote: Edward Martinez wrote on 02/ 4/11 01:52 PM: On 02/04/11 13:06, Bill Shannon wrote: Bart Smaalders wrote on 02/ 4/11 11:49 AM: On 02/04/11 11:33, Bill Shannon wrote: This isn't really specific to OpenSolaris since it also happens on Solaris 10, but maybe someone here can give me some ideas? I have a java program that is failing because it does something that calls fork1(), which fails with ENOMEM (confirmed using truss): 27010/2: fork1() Err#12 ENOMEM I've used both the 32-bit and 64-bit versions of the JVM. Using the 64-bit version, I see the heap expand to just over 4GB before it forks: 27010/30: brk(0x107DE3D90) = 0 I have almost 160 GB of swap space: $ swap -s total: 806336k bytes allocated + 167872k reserved = 974208k used, 159746184k available It doesn't seem like it can possibly be running out of swap space. What other reasons would cause fork1() to fail with ENOMEM? ___ The vm system is rather fond of using ENOMEM as a generic error bucket. If you have a way of reproducing the problem, a bit of DTrace will quickly turn up the reason. I assume it's only the 32 bit version that is failing to fork, right? Sorry, I wasn't clear. The 64-bit version fails as well. With the 32-bit version I limit the Java heap to 2GB. It doesn't run out of heap, but the fork fails. I'm sure dtrace will solve every problem I have, but I don't know how to use it to solve this problem! :-( Hints? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org Hi, Please excuse me for getting into the discussion. I think in solaris 10 the /tmp file is place by default in swap, if /tmp is full it may give that error. I found an article on that theory. http://nilesh-joshi.blogspot.com/2010/03/tmp-file-system-full-swap-space-limit.html /tmp isn't full: $ df -h /tmp Filesystem size used avail capacity Mounted on swap 152G 384K 152G 1% /tmp It's hard to tell if it runs out of space during the execution of the process but I doubt it. I don't get any console messages saying the filesystem is full. Hi, Interesting... it may be running out of heap space, i think limits can be checked with ulimit -a $ ulimit -a time(seconds) unlimited file(blocks) unlimited data(kbytes) unlimited stack(kbytes) 8192 coredump(blocks) unlimited nofiles(descriptors) 256 vmemory(kbytes) unlimited Hi, I think stack may be low. On my solaris 10, defaults are: # ulimit -a time(seconds) unlimited file(blocks) unlimited data(kbytes) unlimited stack(kbytes) 10240 coredump(blocks) unlimited nofiles(descriptors) 256 memory(kbytes) unlimited Why would that cause fork to fail? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] errors from update-refresh.sh cron job
Any idea what causes this? Your cron job on datsunx /usr/lib/update-manager/update-refresh.sh produced the following output: /usr/lib/python2.6/vendor-packages/gtk-2.0/gtk/__init__.py:57: GtkWarning: could not open display warnings.warn(str(e), _gtk.Warning) GConf Error: Failed to contact configuration server; some possible causes are that you need to enable TCP/IP networking for ORBit, or you have stale NFS locks due to a system crash. See http://projects.gnome.org/gconf/ for information. (Details - 1: Not running within active session) /usr/bin/pm-updatemanager:93: GtkWarning: file gtkicontheme.c: line 317: assertion `GDK_IS_SCREEN (screen)' failed self.icon_theme = gtk.icon_theme_get_default() Traceback (most recent call last): File /usr/bin/pm-updatemanager, line 408, in module set_check_all, set_check_cache) File /usr/bin/pm-updatemanager, line 95, in __init__ self.icon_theme.append_search_path(pkg_icon_location) AttributeError: 'NoneType' object has no attribute 'append_search_path' ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] hg status /home?
Anyone have any idea what causes this process to start running? /usr/bin/python2.4 /usr/bin/hg status -marduC --repository /home --cwd /home /home It's very busy accessing all my NFS mounts on /home. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] upgrade woes, snv_132 - snv_147 - snv_149
I upgraded my snv_132 system today. I'm using the Oracle-internal pkg server mirror so I apologize if this is in an Oracle-internal problem. The Update Manager upgraded me to snv_147 even though the latest version of entire available was 149. And it created a new boot environment named opensolaris-133. Ok, weird, but not the real problem. Now, running snv_147, I ran Update Manager and told it to update me again. After gathering the package information it failed with: Error: This is an Live Image. The install operation can't be performed. What does that mean? Oh, and if I click on the Release notes link, it takes me to a document that has been deleted a few days ago. Should I be able to upgrade beyond snv_147? Or should I just wait for Solaris 11 Express? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] Package Manager vs. package repo browser
Why am I seeing different information from Package Manager than what I see when browsing pkg.opensolaris.org/dev? (Yes, Package Manager is pointing at the dev repository.) For instance, if I look at SUNWgmake in Package Manager, I see version 3.81 (Build 5.11-0.133). Browsing the repo I see: sunwgm...@3.81,5.11-0.132:20100130T075514Z What's up with that? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Package Manager vs. package repo browser
Shawn Walker wrote on 02/26/10 04:28 PM: On 02/26/10 06:20 PM, Bill Shannon wrote: Why am I seeing different information from Package Manager than what I see when browsing pkg.opensolaris.org/dev? (Yes, Package Manager is pointing at the dev repository.) For instance, if I look at SUNWgmake in Package Manager, I see version 3.81 (Build 5.11-0.133). Browsing the repo I see: sunwgm...@3.81,5.11-0.132:20100130T075514Z What's up with that? The BUI provides more information than the packagemanager. This isn't an issue of more vs. less, it's an issue of different. One has version 133, one has version 132. Why? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Package Manager vs. package repo browser
Shawn Walker wrote on 02/26/2010 05:44 PM: On 02/26/10 07:30 PM, Bill Shannon wrote: Shawn Walker wrote on 02/26/10 05:05 PM: On 02/26/10 06:49 PM, Bill Shannon wrote: Shawn Walker wrote on 02/26/10 04:28 PM: On 02/26/10 06:20 PM, Bill Shannon wrote: Why am I seeing different information from Package Manager than what I see when browsing pkg.opensolaris.org/dev? (Yes, Package Manager is pointing at the dev repository.) For instance, if I look at SUNWgmake in Package Manager, I see version 3.81 (Build 5.11-0.133). Browsing the repo I see: sunwgm...@3.81,5.11-0.132:20100130T075514Z What's up with that? The BUI provides more information than the packagemanager. This isn't an issue of more vs. less, it's an issue of different. One has version 133, one has version 132. Why? Sorry, I didn't see the 132, but you'd have to be more specific. I see 133 at pkg.opensolaris.org/dev when I browse the packages for build 133: http://pkg.opensolaris.org/dev/en/catalog.shtml?version=0.5.11%2C0.5.11-0.133action=Browse Are you browsing the packages for build 132? No. See attached screenshot. I go to the main screen, search for make, and this is what I see. Because search doesn't do substring matching by default. When you search for make, it's finding version 132 of SUNWgmake because it actually has a file called make in it: basename link usr/gnu/bin/make sunwgm...@3.81,5.11-0.132:20100130T075514Z Since SUNWgmake was renamed to developer/build/gnu-make in build 133, you won't see sunwgm...@133 since it no longer contains any files. Are you comparing search in the packagemanager to search in the BUI? Yup. Shouldn't they do the same thing? Who *wouldn't* expect the gmake/gnu-make package to match the search make? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Package Manager vs. package repo browser
Shawn Walker wrote on 02/26/2010 06:25 PM: On 02/26/10 08:15 PM, Bill Shannon wrote: Shawn Walker wrote on 02/26/2010 05:44 PM: On 02/26/10 07:30 PM, Bill Shannon wrote: Shawn Walker wrote on 02/26/10 05:05 PM: On 02/26/10 06:49 PM, Bill Shannon wrote: Shawn Walker wrote on 02/26/10 04:28 PM: On 02/26/10 06:20 PM, Bill Shannon wrote: Why am I seeing different information from Package Manager than what I see when browsing pkg.opensolaris.org/dev? (Yes, Package Manager is pointing at the dev repository.) For instance, if I look at SUNWgmake in Package Manager, I see version 3.81 (Build 5.11-0.133). Browsing the repo I see: sunwgm...@3.81,5.11-0.132:20100130T075514Z What's up with that? The BUI provides more information than the packagemanager. This isn't an issue of more vs. less, it's an issue of different. One has version 133, one has version 132. Why? Sorry, I didn't see the 132, but you'd have to be more specific. I see 133 at pkg.opensolaris.org/dev when I browse the packages for build 133: http://pkg.opensolaris.org/dev/en/catalog.shtml?version=0.5.11%2C0.5.11-0.133action=Browse Are you browsing the packages for build 132? No. See attached screenshot. I go to the main screen, search for make, and this is what I see. Because search doesn't do substring matching by default. When you search for make, it's finding version 132 of SUNWgmake because it actually has a file called make in it: basename link usr/gnu/bin/make sunwgm...@3.81,5.11-0.132:20100130T075514Z Since SUNWgmake was renamed to developer/build/gnu-make in build 133, you won't see sunwgm...@133 since it no longer contains any files. Are you comparing search in the packagemanager to search in the BUI? Yup. Shouldn't they do the same thing? Who *wouldn't* expect the gmake/gnu-make package to match the search make? The results from a local client may intentionally differ from that of the BUI. For example, the local client may choose to not list SPARC packages when you search, because it knows you're on an x86 system. It might also choose (if you're on b132), to not list packages for b133 in certain cases since you can't install the b133 packages. Local-search filtering is something that is still being explored, so for now, you will continue to see some differences. If you feel that a particular search behaviour isn't what you expect in the client, then feel free to file an RFE or bug against the client in question so that someone can evaluate it. The BUI's behaviour in this case is correct, what each client does is up to them. It's hoped that the advanced search interface the BUI provides will eventually be able to provide you with variant filtering, etc. just as the client does (although it will probably be manual, not automatic). Doesn't sound like you think this is a problem so I won't waste my time and yours by filing a bug. Thanks anyway. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] beadm rename fails
Brian Ruthven - Solaris Network Sustaining - Sun UK wrote on 2/8/10 2:32 AM: Is the BE that you are trying to rename the currently activated BE (that is, active on next reboot)? Yes, it wasn't active since it was newly created, but it was the one that would be activated on the next boot. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] beadm rename fails
Trying to upgrade from b127 to b132. After working around some other bugs, pkg image-update succeeded, but left me with a BE with a stupid name. I tried to rename it (while still running in the old BE) but it fails saying unmount failed. truss shows that it's trying to unmount /! Yow, that's never going to work! Is it just broken, or am I doing something wrong? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] beadm rename fails
Shawn Walker wrote on 02/ 5/10 12:49 PM: On 02/ 5/10 02:39 PM, Bill Shannon wrote: Trying to upgrade from b127 to b132. After working around some other bugs, pkg image-update succeeded, but left me with a BE with a stupid name. I tried to rename it (while still running in the old BE) but it fails saying unmount failed. truss shows that it's trying to unmount /! Yow, that's never going to work! Is it just broken, or am I doing something wrong? I've seen this but don't know the cause, but I wanted to add that you can pick your own BE name during image-update using the --be-name option. Well, I guess I can destroy that BE and start all over again to get a reasonable name. FYI, here's the mounts and unmounts it does: 15197: mount(rpool/ROOT/opensolaris-127-nvidia-1, /tmp/.be.MraORD, MS_OPTIONSTR, zfs, 0x, 0, 0x080451D0, 1024) = 0 15197: mount(rpool/ROOT/opensolaris-127-nvidia-1/opt, /tmp/.be.MraORD/opt, MS_OPTIONSTR, zfs, 0x, 0, 0x08043B50, 1024) = 0 15197: umount2(/tmp/.be.MraORD/opt, 0x)= 0 15197: umount2(/tmp/.be.MraORD, 0x)= 0 15197: umount2(/, 0x) Err#16 EBUSY ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] java classpath
Dr. Martin Mundschenk wrote on 12/ 1/09 02:26 PM: Hi! I'm moving from OSX Server to an OpenSolaris box. There are some java services to install and it's driving me nuts... The are some jar-Files that call classes from the jar file jt400.jar. I always get: Exception in thread main java.lang.NoClassDefFoundError: jt400/jar Caused by: java.lang.ClassNotFoundException: jt400.jar [...] I tried: 1) export CLASSPATH=jt400.jar java -jar SFR.jar 2) java -cp jt400.jar -jar SFR.jar on OSX all I have to do is move the libraries into /Library/Java/Extensions and all works fine. Is there a similar directory in OpenSolaris? CLASSPATH and -cp are ignored when using java -jar. Ideally SFR.jar would have a Class-Path entry in its MANIFEST.MF file referencing jt400.jar. If you know or can figure out the name of the main class in SFR.jar, you can use: java -cp SFR.jar:jt400.jar mainClassName You can also put jt400.jar in the /usr/java/jre/lib/ext directory, if you have permission, or use: java -Djava.ext.dirs=$PWD -jar SFR.jar Hope that helps. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] glissfish
philo neo wrote on 10/1/09 6:36 AM: hi, which port number of glassfish since fire fox. htt://127.0.0.1: By default the admin port is 4848 and the application port is 8080. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] glissfish
philo neo wrote on 10/1/09 12:22 PM: i try 4848 and 8080 with https and http: no reponses !! can i to reinstall glassfish or it's a another probleme? Probably best to take your question to the GlassFish user's forum or mailing list for debugging help. us...@glassfish.dev.java.net https://glassfish.dev.java.net/servlets/ProjectMailingListList http://forums.java.net/jive/forum.jspa?forumID=56start=0 ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] OT: Sun keyboard on Ubuntu Linux 8.04
This is a bit off topic, but... I thought I remembered a discussion here that would've helped me figure this out, but I can't find it... I'm trying to get a Sun keyboard to work with Ubuntu. Everything works except the left keypad. I really want Cut/Copy/Paste, and I'd like Front to work. I can't figure out what to configure in Gnome to enable these to work. I was hoping to see what was done in OpenSolaris to make them work and then copy that for Ubuntu. Can anyone point me in the right direction? Thanks! ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Login Delay
SOBRAL Thiago wrote: Hi everyone, Has anyone experienced a long delay to login into OpenSolaris? The login can be local (console), ssh or telnet. The time that I’m talking about is around 10 minutes. I checked the DNS entries resolv.conf / nsswitch.conf and everything is oK. Do you have any idea why this system can take this behavior ? I've experienced login delays with both OpenSolaris and Solaris for years. The first time after rebooting it takes several minutes to login. Since I only reboot once a year or less, I haven't taken the time to track it down. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] RES: Login Delay
As I said, it's the reboot that *causes* my problem. The first login after reboot is slow. Logins after that work normally. SOBRAL Thiago wrote: Hi Bill, I've the same problem with Solaris10 into V240 box. When I try to login, I take a coffee and wait 10 long minutes to log in. Any idea without a reboot ? Thiago Sobral -Mensagem original- De: Bill Shannon [mailto:bill.shan...@sun.com] Enviada em: quinta-feira, 20 de agosto de 2009 17:29 Para: SOBRAL Thiago Cc: opensolaris-discuss@opensolaris.org Assunto: Re: Login Delay SOBRAL Thiago wrote: Hi everyone, Has anyone experienced a long delay to login into OpenSolaris? The login can be local (console), ssh or telnet. The time that I'm talking about is around 10 minutes. I checked the DNS entries resolv.conf / nsswitch.conf and everything is oK. Do you have any idea why this system can take this behavior ? I've experienced login delays with both OpenSolaris and Solaris for years. The first time after rebooting it takes several minutes to login. Since I only reboot once a year or less, I haven't taken the time to track it down. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] OpenSolaris development package repository build 111a, x86/x64/SPARC
david.co...@sun.com wrote: Other known issues in this repository update 6811922 PCIE init err info failed BDF 0x0 on snv_109 (Ultra 20 M2) I asked about this before but never got a complete answer. Let's try again... I have an Ultra 20, not an Ultra 20 M2, does that mean this doesn't apply? I believe the issue affects both the Ultra 20 and the Ultra 20 M2. Thanks. Maybe someone could update the description? If it does apply, can someone tell me how to determine what my firmware version is, and how to upgrade it if it's too old? I found the download referenced above, but it comes with no instructions at all, and a web search didn't turn up anything. The easiest way to find out the firmware version on that platform would be to run /usr/sbin/smbios and look for the SMB_TYPE_BIOS section. I tried that. It didn't work: $ smbios smbios: failed to load SMBIOS: System does not export an SMBIOS table There should be a Version String field. Alternatively, the BIOS should report its version when the system is powered on. For the Ultra 20, it looks like the Sun Ultra 20 Workstation Supplemental 1.8 ISO Image contains the 2.3.0 BIOS which can be burned on a CD and then booted. It should prompt you at that point with a menu which will flash the new BIOS into place. I ended up doing that. It then complained about a CMOS checksum error or something like that, so I had to do Load Optimized Settings, which made it happy. After that, I did the upgrade, which led to a whole other set of problems in a following message... ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] snv_111a kernel crashes at boot
I just upgraded (using package manager upgrade all) from snv_101a to snv_111a. When I reboot in the new boot environment, it crashes right away. Since it reboots after the crash, clearing the screen, it's really hard to catch why it's failing. I resorted to taking a picture of the screen before the reboot occurs. Surely there's a better way? Anyway, it dies with some sort of panic. The stack trace looks like: unix:mutex_panic unix:mutex_destroy gfx_private:gfxp_vga... atiatom:atiatom_detac... atiatom:atiatom_attac... genunix:devi_atach genunix:attach_node genunix:i_ndi_config_... genunix:i_ddi_attachc... genunix:i_ddi_attach_... consconfig_dacf:plat_... consconfig_dacf:consol... It's an Ultra 20, with two screens. Any idea what's going wrong? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] snv_111a kernel crashes at boot
John Martin wrote: Bill Shannon wrote: I just upgraded (using package manager upgrade all) from snv_101a to snv_111a. When I reboot in the new boot environment, it crashes right away. Since it reboots after the crash, clearing the screen, it's really hard to catch why it's failing. I resorted to taking a picture of the screen before the reboot occurs. Surely there's a better way? Anyway, it dies with some sort of panic. The stack trace looks like: unix:mutex_panic unix:mutex_destroy gfx_private:gfxp_vga... atiatom:atiatom_detac... atiatom:atiatom_attac... genunix:devi_atach genunix:attach_node genunix:i_ndi_config_... genunix:i_ddi_attachc... genunix:i_ddi_attach_... consconfig_dacf:plat_... consconfig_dacf:consol... It looks like gfxp_private::gfxp_vgatext_detatch() is being called twice, once when gfx_private::gfxp_vgatext_attach() detects an error and cleans up after itself and again after it returns the DDI_FAILURE error code to the atiatom module and it calls gfxp_vgatext_detach() again from atiatom_detach(). mutex_destroy() gets called twice for the same object. File a bug. The gfx_private driver needs to keep state on whether the lock needs to be destroyed in case its detach is called multiple times. However, the bigger problem will eventually be why the attach function failed since all it does is basic aperture mapping. Filing a bug for this is a waste of my time. If no one else is running into this, it's not going to be fixed anytime soon. And if other people *are* running into it, it will be fixed without *me* filing a bug. It's less frustrating to just hope that the next build will be better. At least, that's been my experience so far. What I'm hoping for is some sort of workaround... It's an Ultra 20, with two screens. Is this Ultra 20 or Ultra 20 M2? If M2, have you updated the most recent SBIOS? Not M2. I upgraded the BIOS anyway, just to avoid that excuse. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] OpenSolaris development package repository build 111a, x86/x64/SPARC
david.co...@sun.com wrote: Other known issues in this repository update 6811922 PCIE init err info failed BDF 0x0 on snv_109 (Ultra 20 M2) Ultra 20 M2 systems may not boot after being upgraded or installed with build 109. Work-around: Upgrade the system's firmware to at least revision 1.52. Updated firmware can be obtained from http://www.sun.com/desktop/workstation/ultra20/downloads.jsp I asked about this before but never got a complete answer. Let's try again... I have an Ultra 20, not an Ultra 20 M2, does that mean this doesn't apply? If it does apply, can someone tell me how to determine what my firmware version is, and how to upgrade it if it's too old? I found the download referenced above, but it comes with no instructions at all, and a web search didn't turn up anything. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] upgrading Ultra 20 to snv_111
In the announcement of snv_111 it says: 6811922 PCIE init err info failed BDF 0x0 on snv_109 (Ultra 20 M2) Ultra 20 M2 systems may not boot after being upgraded or installed with build 109. Work-around: Upgrade the system's firmware to at least revision 1.52. Updated firmware can be obtained from http://www.sun.com/desktop/workstation/ultra20/downloads.jsp I have an Ultra 20, not an Ultra 20 M2, does that mean this doesn't apply? If it does apply, can someone tell me how to determine what my firmware version is, and how to upgrade it if it's too old? I found the download referenced above, but it comes with no instructions at all, and a web search didn't turn up anything. Thanks. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] kernel panic - #df Double fault
My OpenSolaris 101a server at home has died just after 3:00am two times, both with the same panic (included below). In addition to the standard root crontab entries that run at 3:00am: 10 3 * * * /usr/sbin/logadm 15 3 * * 0 [ -x /usr/lib/fs/nfs/nfsfind ] /usr/lib/fs/nfs/nfsfind 30 3 * * * [ -x /usr/lib/gss/gsscred_clean ] /usr/lib/gss/gsscred_clean I also have a crontab entry that uses smbfs to access another machine and uses rdist to copy files from that other machine to the same machine (using rdist in pull instead of push mode, effectively). My kernel crash dump debugging skills are rusty. Can someone help me interpret the information from the panic to determine what's causing the problem? Thanks. Dec 6 03:16:11 nissan unix: [ID 836849 kern.notice] Dec 6 03:16:11 nissan ^Mpanic[cpu0]/thread=d3112de0: Dec 6 03:16:11 nissan genunix: [ID 683410 kern.notice] BAD TRAP: type=8 (#df Do uble fault) rp=fec24128 addr=0 Dec 6 03:16:11 nissan unix: [ID 10 kern.notice] Dec 6 03:16:11 nissan unix: [ID 839527 kern.notice] sched: Dec 6 03:16:11 nissan unix: [ID 753105 kern.notice] #df Double fault Dec 6 03:16:11 nissan unix: [ID 243837 kern.notice] pid=0, pc=0xfebe6d8e, sp=0x d3110ffc, eflags=0x10296 Dec 6 03:16:11 nissan unix: [ID 211416 kern.notice] cr0: 8005003bpg,wp,ne,et,t s,mp,pe cr4: 6d8xmme,fxsr,pge,mce,pse,de Dec 6 03:16:11 nissan unix: [ID 624947 kern.notice] cr2: d3110ff8 Dec 6 03:16:11 nissan unix: [ID 625075 kern.notice] cr3: 25dd000 Dec 6 03:16:11 nissan unix: [ID 10 kern.notice] Dec 6 03:16:11 nissan unix: [ID 537610 kern.notice] gs: 1b0 fs: 0 es: 160 ds: 160 Dec 6 03:16:11 nissan unix: [ID 537610 kern.notice]edi:0 esi: 0 ebp:0 esp: fec24160 Dec 6 03:16:11 nissan unix: [ID 537610 kern.notice]ebx: d8d19254 edx: 0 ecx:0 eax: f94553b4 Dec 6 03:16:11 nissan unix: [ID 537610 kern.notice]trp:8 err: 0 eip: febe6d8e cs: 158 Dec 6 03:16:11 nissan unix: [ID 717149 kern.notice]efl:10296 usp: d3110 ffc ss: 160 Dec 6 03:16:11 nissan unix: [ID 125119 kern.notice] tss.tss_link: 0x0 Dec 6 03:16:11 nissan unix: [ID 125119 kern.notice] tss.tss_esp0: 0xd2befe 34 Dec 6 03:16:11 nissan unix: [ID 125119 kern.notice] tss.tss_ss0: 0x160 Dec 6 03:16:11 nissan unix: [ID 125119 kern.notice] tss.tss_esp1: 0x0 Dec 6 03:16:11 nissan unix: [ID 125119 kern.notice] tss.tss_ss1: 0x0 Dec 6 03:16:11 nissan unix: [ID 125119 kern.notice] tss.tss_esp2: 0x0 Dec 6 03:16:11 nissan unix: [ID 125119 kern.notice] tss.tss_ss2: 0x0 Dec 6 03:16:11 nissan unix: [ID 125119 kern.notice] tss.tss_cr3: 0x25dd00 0 Dec 6 03:16:11 nissan unix: [ID 125119 kern.notice] tss.tss_eip: 0xfebe6d 8e Dec 6 03:16:11 nissan unix: [ID 125119 kern.notice] tss.tss_eflags:0x10296 Dec 6 03:16:11 nissan unix: [ID 125119 kern.notice] tss.tss_eax: 0xf94553 b4 Dec 6 03:16:11 nissan unix: [ID 125119 kern.notice] tss.tss_ebx: 0xd8d192 54 Dec 6 03:16:11 nissan unix: [ID 125119 kern.notice] tss.tss_ecx: 0x0 Dec 6 03:16:11 nissan unix: [ID 125119 kern.notice] tss.tss_edx: 0x0 Dec 6 03:16:11 nissan unix: [ID 125119 kern.notice] tss.tss_esp: 0xd3110f fc Dec 6 03:16:11 nissan unix: [ID 10 kern.notice] Dec 6 03:16:11 nissan last message repeated 1 time Dec 6 03:16:11 nissan genunix: [ID 672855 kern.notice] syncing file systems... Dec 6 03:16:11 nissan genunix: [ID 904073 kern.notice] done Dec 6 03:16:12 nissan genunix: [ID 111219 kern.notice] dumping to /dev/zvol/dsk /rpool/dump, offset 65536, content: kernel Dec 6 03:16:28 nissan genunix: [ID 409368 kern.notice] ^M100% done: 115151 page s dumped, compression ratio 1.87, Dec 6 03:16:28 nissan genunix: [ID 851671 kern.notice] dump succeeded ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] kernel panic - #df Double fault
Sebastien Roy wrote: On Sat, 2008-12-06 at 10:48 -0800, Bill Shannon wrote: Dec 6 03:16:11 nissan unix: [ID 836849 kern.notice] Dec 6 03:16:11 nissan ^Mpanic[cpu0]/thread=d3112de0: Dec 6 03:16:11 nissan genunix: [ID 683410 kern.notice] BAD TRAP: type=8 (#df Do uble fault) rp=fec24128 addr=0 A double fault on 32-bit x86 is usually indicative of a blown stack. This could be a bug due to infinite recursion, or simply a very deep stack. Sherry Moore has increased the default stack size on 32-bit from 2 to 3 pages in build 105, and that will hopefully alleviate instances of the latter. See: 6766854 BAD TRAP: type=8 (#df Double fault) (Xeon) For some immediate relief (assuming that you're not running into some infinite recursion bug somewhere), you can increase your stack size by placing the following in /etc/system (a reboot is required for this to take effect): set _defaultstksz=0x5000 Thanks, I'll try that. Assuming it's not a recursion bug, is there any interest in finding out what's blowing the stack? Probably not if you've already decided to increase the default stack size. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] pkg.opensolaris.org is down
pkg.opensolaris.org seems to be down. Is this on purpose? Did I miss the announcement? Isn't this site run on our high availability infrastructure? Is there another site, perhaps an internal Sun site, that I can use to upgrade to the latest build? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] pkg.opensolaris.org is down
Bill Shannon wrote: pkg.opensolaris.org seems to be down. Is this on purpose? Did I miss the announcement? Isn't this site run on our high availability infrastructure? Is there another site, perhaps an internal Sun site, that I can use to upgrade to the latest build? Ok, maybe it's not completely down, but it's so slow as to be unusable. Fortunately, I found the SWAN instructions: http://infoshare.sfbay.sun.com/twiki/bin/view/Pkg/ ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] upgrade from 101a to 101b?
The instructions on http://opensolaris.org/os/project/indiana/resources/relnotes/200805/image-update/ don't work for a build number of 101a. How do I upgrade from 101a to 101b? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] guest access to smb/cifs share?
How do I enable guest (unauthenticated) access to a smb/cifs share using the zfs set command? This page: http://blogs.sun.com/timthomas/entry/solaris_cifs_in_workgroup_mode claims it isn't supported, but hopefully that's old data. I've read man pages, admin manuals, and googled, but I can't figure this out. Oh, I'm using OpenSolaris 2008.11 101a. Thanks. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] Error 16: Inconsistent filesystem structure
I just upgraded to snv_101a_rc1b. I rebooted in the new environment and everything seemed fine. Then I rebooted again and I got the error Error 16: Inconsistent filesystem structure and it wouldn't boot. I booted into the previous boot environment and that worked fine. But when I rebooted again and tried to boot into either the new or the previous boot environment, I got the same error. So I rebooted into the boot environment before the previous boot environment and that worked. Now I'm scared to reboot at all because I'm out of boot environments. A few web searches turned up other people with the same problem, but no obvious solutions. For example: http://opensolaris.org/jive/thread.jspa?threadID=80625 http://opensolaris.org/jive/thread.jspa?messageID=303905 http://defect.opensolaris.org/bz/show_bug.cgi?id=3965 Can someone please tell me how to fix my boot environment so that it will work? I *really* don't want to have to reinstall this system from scratch. Thanks! ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] crontab
Johan Hartzenberg wrote: On Fri, Oct 17, 2008 at 9:43 PM, Bill Shannon [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] wrote: Johan Hartzenberg wrote: cediag will put its results into the messages file if scheduled, or print a report on the console if executed manually. Whether you run it manually or scheduled via cron you in any case have to monitor the messages file, as you should be doing as part of best practices in any case. Um, what's the current best practice on how to do that? I said many things in that paragraph. Ya, sorry, all I was asking about is, in general, best practices for monitoring the messages file. I wasn't asking specifically about cediag. The best practive on Monitoring the messages file is to have operators working shifts with consoles which show alerts as they happen and ideally some mechanism to assist them in knowing which messages are meaningful and which ones requires action. In addition, you should have documented daily maintenance/check procedures for Sys Admins, part of which should be to review messages in log files. Kind of excessive for my desktop machine! In the olden days I could start up a console window that would display all the messages logged to the console. Now the messages just go into the log file. You still can do this. /etc/syslog.conf controls which messages goes to the console, which is generally very little these days. To test: open a console window: /usr/dt/bin/dtterm -C Then produce a test alert: logger -p user.alert THIS IS A TEST. This should show up in the above dtterm window. Isn't dtterm, like, obsolete? Can gnome-terminal do this? Is there a tool to display messages from the log file and alert me when there's an issue I should be paying attention to? There are many such tools. Look toward BMC Patrol, CA Unicenter/TNG, etc. I'm looking for a simple single user desktop tool. Ideally, something that sat in the system tray / notification area of the Gnome desktop. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] crontab
Johan Hartzenberg wrote: On Sat, Oct 18, 2008 at 9:19 AM, Bill Shannon [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] wrote: Is there a tool to display messages from the log file and alert me when there's an issue I should be paying attention to? There are many such tools. Look toward BMC Patrol, CA Unicenter/TNG, etc. I'm looking for a simple single user desktop tool. Ideally, something that sat in the system tray / notification area of the Gnome desktop. I think you should research this and select the solution which matches your budget (disk space, CPU consumed, and money spent, as well as the amount of time you want to spend administering the solution) You should also find out why cediag is causing your system to run slowly. Do you have exceptionally large /var/adm/messages files? *I'm* not the one with the cediag problem. That's someone else. I'm just asking about monitoring the messages file. I'm researching it by asking here. My budget is $0. I'm hoping there's an obvious solution that's part of OpenSolaris. I found System Log Viewer, although it gives lots of error messages when started up. I guess that's the right answer once it works. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] stunnel without SSL
Boyd Adamson wrote: Bill Shannon [EMAIL PROTECTED] writes: Anyone know if it's possible to use stunnel without the s part? That is, I just want a tunnel, no SSL. Can I configure stunnel to just tunnel traffic on port X to a remote machine's port X, without injecting SSL? Sounds like you want netcat (nc(1))[1] or (my preference) socat[2] nc did the trick, thanks! ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] stunnel without SSL
Anyone know if it's possible to use stunnel without the s part? That is, I just want a tunnel, no SSL. Can I configure stunnel to just tunnel traffic on port X to a remote machine's port X, without injecting SSL? Thanks. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] network configuration problem
James Carlson wrote: Is there still a problem here in the networking component itself (other than the GUI)? If so, I'd like to help solve it. No, after configuring it all by hand, networking is working fine. The question is, why didn't the GUI find the interface? Is it something about this interface on this machine? Or is this version of the GUI broken? I know, this is not the place to ask these questions. As for the complexity, this is why we strongly recommend using DHCP for configuration. That's what it's designed to do, and it makes things much simpler. Kind of sucks for an NFS server to change its IP address on every reboot. Yes, I know, with sufficient cleverness and control over the DHCP server, I can get it to hand out the same IP address every time. I find it's actually easier to configure the server itself. And it's worked on all previous OpenSolaris releases I've used. It should not take much cleverness to make this happen. Unless the DHCP server is broken or the address space is overcommitted, you should get the same IP address every single time. That's by design. That hasn't been my experience. It often happens, but not always. Maybe it only changes if the machine has been down for awhile. And if you really want a specific one, then nailing that address to a specific MAC address or client ID is pretty easy with most decent servers. It's usually just one command. I was able to do that with a newer Linksys router, but the one I'm using now is older and it doesn't seem to have that capability. And, in fact, it's not even necessary to get addresses from DHCP; it's possible to get just the configuration parameters there. But that is a fair bit harder to set up. In any event, once you've disabled the 'nwam' service and reenabled the 'physical:default' one, the rest of the networking system works exactly as it has for many, many years on Solaris. Thus, if you're running into trouble here, and if you're already familiar with that system (enough so that you consider it easier), then it sounds like you've encountered a bug of some sort. It'd be good to get some actual command output and details so that we can narrow down the cause of the bug. I'm waiting for feedback from the GUI people. I'm not convinced that command output from a correctly working (except GUI) system is going to be helpful, but here's the ifconfig output: $ ifconfig -a lo0: flags=2001000849UP,LOOPBACK,RUNNING,MULTICAST,IPv4,VIRTUAL mtu 8232 index 1 inet 127.0.0.1 netmask ff00 iprb0: flags=201000843UP,BROADCAST,RUNNING,MULTICAST,IPv4,CoS mtu 1500 index 2 inet 192.168.1.10 netmask ff00 broadcast 192.168.1.255 lo0: flags=2002000849UP,LOOPBACK,RUNNING,MULTICAST,IPv6,VIRTUAL mtu 8252 index 1 inet6 ::1/128 ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] network configuration problem
James Carlson wrote: I am *so* sorry I upgrade to snv_98... I don't see how that relates to the previous poster's note about snv_100. I had snv_86 working quite well on two machines and I was very happy with it. Since upgrading those machines and installing snv_98 on another machine (the one that started this thread), I've found quite a few things that are worse than snv_86. When I point out the problems, they're all expected to be fixed in snv_100 or snv_101 or snv_102. I just feel like I upgraded too soon. If you had a working *static* configuration at some point, and an upgrade broke it No, that's not it. As for the static network configuration logic, there's nobody working on that, so it's static. And the GUI is broken. I just tried the GUI on another of my snv_98 machines and it seems to be working fine. At least, it finds the interface. So maybe it's something peculiar about the original machine or the interface on that machine that's causing the problem? You don't suppose it's confused because the interface name is 4 characters (iprb) instead of three (nge)? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] network configuration problem
James Carlson wrote: Looking through the truss output on /usr/bin/network-admin, it appears that it invokes '/usr/sbin/dladm show-link -p'. I think that means it may have been broken by the integration of CR 6722523 in snv_96. I don't know why that hasn't been addressed in GNOME. (The output format of that command changed incompatibly in that fix.) Kludging dladm to produce the old format output didn't fix the problem. (Maybe I did it wrong.) truss also shows that network-admin tries to run a command named -1, even before it runs dladm. That suggests a problem somewhere else. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] network configuration problem
James Carlson wrote: Bill Shannon writes: I disabled svc:/network/physical:nwam and enabled svc:/network/physical:default. I went to System Administration Network. It doesn't show *any* interfaces. I'd recommend contacting the Desktop community. They're the ones who support the GUI bits. I can't say I use them much myself. This was a problem 15 years ago when I left the OS group and I see it's still a problem today. You OS guys need to understand that most people experience desktop systems through the GUI. If the GUI for *your* feature doesn't work, it's the same as your feature not working. Let the flaming begin! I rebooted. No help. I suggest diagnosing the problem before going through repeated reboots. The system is meant to be deterministic: if you've got a problem now, you'll very likely still have that problem after a reboot. Yes, most of the time that's true. But I learned long ago that it's better to reboot than to spend hours and hours trying to diagnose such problems. ifconfig -a looks fine, but there's no default route. What does looks fine mean? Please provide the command output. You know, UP, correct IP address, etc. It's working. Really. What does dladm show-link say? $ dladm show-link LINKCLASSMTUSTATEOVER iprb0 phys 1500 up -- Normal, right? Ok, I can do route add default 192.168.1.1 and now I have a default route. Do this: # echo 192.168.1.1 /etc/defaultrouter Yes, I forgot to mention that I knew how to do that too. But I can't lookup internet addresses, so I copy nsswitch.dns to nsswitch.conf and that works. You'll also want to set up /etc/resolv.conf. It had already been set up, probably by NWAM. But I'm putting all this together by hand, and I never had to do this on any of the other OpenSolaris machines I've installed. Why isn't the Network admin GUI working to allow me to configure the network? No idea; contact the desktop group. Will do. As for the complexity, this is why we strongly recommend using DHCP for configuration. That's what it's designed to do, and it makes things much simpler. Kind of sucks for an NFS server to change its IP address on every reboot. Yes, I know, with sufficient cleverness and control over the DHCP server, I can get it to hand out the same IP address every time. I find it's actually easier to configure the server itself. And it's worked on all previous OpenSolaris releases I've used. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] network configuration problem
Ghee Teo wrote: The new network admin GUI is all integrated to work to its full functionality in snv 100 as far as I heard. What exactly does that mean? The perfectly functional version in older releases was replaced with a new partially functional version? I am *so* sorry I upgrade to snv_98... ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] network configuration problem
I installed 2008.05 on a new machine, upgraded to snv_98. Now I'm trying to convert to static IP. I disabled svc:/network/physical:nwam and enabled svc:/network/physical:default. I went to System Administration Network. It doesn't show *any* interfaces. Sigh. I rebooted. No help. I created /etc/hostname.iprb0 with my IP address. I rebooted. No help. ifconfig -a looks fine, but there's no default route. Ok, I can do route add default 192.168.1.1 and now I have a default route. But I can't lookup internet addresses, so I copy nsswitch.dns to nsswitch.conf and that works. But I'm putting all this together by hand, and I never had to do this on any of the other OpenSolaris machines I've installed. Why isn't the Network admin GUI working to allow me to configure the network? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] printing from firefox3 vs firefox2...
Norm Jacobs wrote: Other than the fact that firefox dumps core when you print, it seems to print. You may very well be running into 6743432 firefox is corrupting it's malloc buffers. which appears to be the cause of 6750542 memory corruption in firefox breaks printing It looks like LD_PRELOAD of libc.so before running firefox works around this. But, as I said on desktop-discuss... Now the only problem is that the printout is completely wrong! It's using the wrong fonts with the wrong size and the wrong spacing! Not just a little bit wrong, but *completely* wrong. The header and footer titles are actually correct, but the font in the page is either squished together so close that the characters overlap each other, or it's half the size it should be and spaced twice as far apart as it should be so that it takes up the same amount of horizontal space. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] irefox 3.0.1 contrib. builds for Solaris10 hungs when printing is attempted
Thommy M. Malmström wrote: I would like to mention one problem though relating to printing: as soon as I use CtrP or use the menus to print a web page, FF3.0.1 FREEZES UP completely. This is the exact same behavior that I see on OpenSolaris 2008.11 b96 It doesn't hang for me, it sort of dies. All the windows disappear and when I get tired of waiting for it to completely die I kill it so I can restart it. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] snv_98 iso
Dave Uhring wrote: Already did that. It helps a little but it's still unusable. Same goes for thunderbird. It's the whole window system that is malfunctioning. Switching between workspaces is a 5 second operation. Machine is Ferrari 4000 with 2GB and fresh install of 2008.11 b86 Then upgrade your system to b98 and replace firefox as I previously showed. I have no idea what kind of processor or graphics your Ferrari uses but the nVidia drivers have been updated considerably since b86 was released. This system works almost perfectly: $ head -1 /etc/release OpenSolaris 2008.11 snv_98 X86 Download the iso from http://www.genunix.org On a different but related topic... I downloaded the snv_98 iso, burned it to DVD (it wouldn't fit on a CD), but when I try to boot on an old Dell 4550 it fails in a different way each time I try - sometimes disk errors, sometimes some problem with some service, sometimes other things, all really weird failures. An old 2008.05 CD that I burned boots and runs just fine. Did I make a mistake by burning it to DVD? I'm going to try installing 2008.05 and upgrade my way to snv_98. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] snv_98 iso
Dave Uhring wrote: I downloaded the snv_98 iso, burned it to DVD (it wouldn't fit on a CD), The osol-0811-98-global.iso is ~661 MB in size. Not the one I downloaded: $ ls -l osol-0811-98.iso -rw-rw-r-- 1 shannon uucp 783020032 2008-09-20 17:13 osol-0811-98.iso I didn't think I need support for all languages. Was that my mistake? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] firefox problems with snv_98
Ginn Chen wrote: It would be nice if there was more information than it's broken. I can't tell if the problems I'm having are the same problems reported above. Firefox 3 in snv_97/snv_98 has several issues. e.g. Bookmarks are missing. Bookmarks could not be saved. Bookmarks could not be renamed. Can't import/export bookmarks. Firefox coredumps at startup. Firefox couldn't work with libumem.so, I'm curious... What kind of testing was done before release that none of these problems were discovered? I would think the first one would've been discovered immediately had anyone actually used it before release. And what turned out to be the root cause of these problems? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] firefox problems with snv_98
Andras Barna wrote: Firefox 3 in u97 is broken. You can download a contribute Firefox 3.0.1 on mozilla.org as a workaround. http://releases.mozilla.org/pub/mozilla.org/firefox/releases/3.0.1/contrib/solaris_tarball/ It will be fixed in u99. http://mail.opensolaris.org/pipermail/desktop-discuss/2008-September/015699.html Thanks, I'll try that. It would be nice if there was more information than it's broken. I can't tell if the problems I'm having are the same problems reported above. BTW, where does stdout/stderr of programs launched from the panel go? I don't see anything related to firefox in .xsession-errors, for instance. But I do see a long stream of no printer added since last 60 secs which seems to come from http://hg.opensolaris.org/sc/src/presto/ospm/trunk/applet/ospm-hal-support.c Another bug? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] no printer added since last 60 secs
Evan Yan wrote: Bill Shannon wrote: BTW, where does stdout/stderr of programs launched from the panel go? I don't see anything related to firefox in .xsession-errors, for instance. But I do see a long stream of no printer added since last 60 secs which seems to come from http://hg.opensolaris.org/sc/src/presto/ospm/trunk/applet/ospm-hal-support.c Another bug? Hi Bill, The messages are logged by ospm-applet, the applet for Solaris Print Manager[1]. It's not an error, but just some log messages. A message every minute to say things are normal seems pretty gratuitous, and it's going to fill up my .xsession-errors file. Shouldn't this only occur if some extra logging or debugging is enabled? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] firefox problems with snv_98
I just updated my OpenSolaris machine to snv_98. Sure is nice to see firefox 3.0. Unfortunately, it lost all of my bookmarks. I tried to restore them several ways but nothing worked. It seems that bookmarks are now only stored as json. I exported the bookmarks from another machine and imported them as json and that worked. Still, why were they lost to begin with? I use Foxmarks to synchronize my bookmarks between the 4 or more machines I use. Foxmarks wasn't able to restore my bookmarks from the server. It got stuck in loading state file or something like that and never completed. Now that I've restored my bookmarks manually, I'm trying to use Foxmarks to resynchronize my bookmarks. It's stuck in downloading sync file. Anyone have any idea what's going on with this latest version of firefox? Is there an easy way to revert to the previous version of firefox? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] firefox problems with snv_98
Brent Jones wrote: Can you export your previous bookmarks into HTML format, then reimport them that way? No. I do the import, but they don't show up. As I said, I finally was able to import the bookmarks, but... 1. Why did they disappear to begin with? Anyone else seen this? Is it a bug? 2. Why doesn't Foxmarks work anymore? (Ditto) ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] Lightning alarms don't work
Anyone else using Lightning (the Thunderbird calendar plugin)? I was happy when updating to snv_91 got me the most recent version, but still I can't get alarms to work. I've got alarms turned on, both sound and alarm box, and my appointments have alarms set for 5 minutes before the appointment, but still I get nothing. No sound, no alarm box, nothing. Is this broken or am I doing something wrong? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Lightning alarms don't work
Sergio Enrique Schvezov wrote: Bill Shannon wrote: Anyone else using Lightning (the Thunderbird calendar plugin)? I was happy when updating to snv_91 got me the most recent version, but still I can't get alarms to work. I've got alarms turned on, both sound and alarm box, and my appointments have alarms set for 5 minutes before the appointment, but still I get nothing. No sound, no alarm box, nothing. I'm on indy_91 and am using lightning with the google calendar provider addon and am getting notifications (popups + console beeps) for my google calendar events I'm using the Sun calendar server. Further investigation indicates that it's a bug in the Sun calendar server where it doesn't properly store the attributes that Lightning uses for the alarms. There's a fix available, but it hasn't been deployed within Sun yet. Sigh... ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Lock Screen menu item?
Matt Keenan wrote: Do you have any of the lockdown features enabled ? 1. /desktop/gnome/lockdown/restrict_application_launching ? if this is set then ensure /usr/X11/bin/xscreensaver and /usr/X11/bin/xscreensaver-demo are both specified in the list of allowed applications gconf key : /desktop/gnome/lockdown/allowed_applications if this is not set then ignore allowed applications None of the boxes is checked, but there's a long list of applications in allowed_applications. Can I just put * in the list? The list includes /usr/bin/xscreensaver-demo. There's no /usr/X11/bin commands in the list at all. The same is true on the account where Lock Screen shows up in the menu. Where does this list come from? I can't find it in any of the dot files in my home directory. I guess it must be coming from some system location. Even though I don't think it's being used here (because restrict_application_launching isn't set), the list is still wrong. 2. The menu item will also not appear if your session if gnome-panel cannot find the binaries so is /usr/X11/bin in your PATH ? No, but /usr/openwin/bin is (historical artifact) and /usr/openwin is a symlink to /usr/X11. Ok, after a bunch of trial and error, I figured out the problem. I was only adding /usr/openwin/bin to my PATH if $GNOME_DESKTOP_SESSION_ID was set. Otherwise, I assumed I wasn't logging in in a window environment. Unfortunately, that doesn't seem to be set when the login process first runs my .profile, even though it *is* set when I start gnome-terminal. Thanks for the hint that pointed me in the right direction! ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] OpenOffice font problem
W. Wayne Liauh wrote: Again, for and only for the purpose of testing (b/c of potential licensing violations), you can copy the truetype fonts from /usr/openwin/lib/X11/fonts/TrueType in SXDE into a USB stick, then copy them to ~/.fonts of your home directory. You may also need run the fc-cache command before you can use these fonts. Yup, that solved my problem. Thanks! Sun sells StarOffice8 in either download or boxed form at a very reasonable price: http://www.sun.com/software/star/staroffice/get.jsp The same license number can be used in Linux, Solaris, and Windows. Except it doesn't include the Arial Narrow font that's causing my problems. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] OpenOffice font problem
W. Wayne Liauh wrote: This may be a Sun-specific problem, but... Now that I'm running OpenSolaris on my Ultra 20 desktop instead of SXDE on my SunBlade 2500, I'm using OpenOffice instead of StarOffice. Unfortunately, many fonts display differently. In particular, fonts used widely in Sun presentations, such as Arial Narrow. This causes presentation text to no longer fit on the slide. Has anyone figured out how to get these fonts to display the same way they do in StarOffice? (BTW, the same thing has happened for years on my Linux system. In an act of desperation, I even installed StarOffice on my Linux machine, but that didn't solve the problem. Possibly it's an Xorg vs. Xsun issue, although other people claim that installing StarOffice fixes the problem for them.) ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org You may want to try to copy your fonts from SXDE to 2008.05 to see if this solves your problem. Plse note that due to licensing limitations, you may not be able to use some of those fonts other than for testing purposes. Would installing the SUNWxwcft package from SXDE be sufficient? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Lock Screen menu item?
Matt Keenan wrote: Bill, You may have the disable_lock_screen gconf key set, using gconf-editor check : /apps/panel/global/disable_lock_screen and ensure it is not set. It is not set. Other ideas? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] OpenOffice font problem
This may be a Sun-specific problem, but... Now that I'm running OpenSolaris on my Ultra 20 desktop instead of SXDE on my SunBlade 2500, I'm using OpenOffice instead of StarOffice. Unfortunately, many fonts display differently. In particular, fonts used widely in Sun presentations, such as Arial Narrow. This causes presentation text to no longer fit on the slide. Has anyone figured out how to get these fonts to display the same way they do in StarOffice? (BTW, the same thing has happened for years on my Linux system. In an act of desperation, I even installed StarOffice on my Linux machine, but that didn't solve the problem. Possibly it's an Xorg vs. Xsun issue, although other people claim that installing StarOffice fixes the problem for them.) ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] Lock Screen menu item?
Anyone know how to add the Lock Screen menu item to a menu? I know it exists in the default menu but I long ago customized the default menu and my customized version doesn't have it. It's not in my Add to Panel dialog, but if I log in as a user using the default configuration, it does appear in that Add to Panel dialog. I'm sure there's something about my customized configuration that's causing it to not show up. Any idea how to get Lock Screen back? Also, how do I customize the things under the System menu? The Main Menu preferences dialog doesn't show most of the items on the System menu, and doesn't seem to let me add new items to the System menu. BTW, I'm using a non-default theme, if that matters. Thanks for any help! ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] screen resolution problem with OpenSolaris 2008.05 on VirtualBox 1.6
Bill Shannon wrote: I spent several hours trying to get OpenSolaris 2008.05 (final version) to work with a screen resolution larger than 1024x768 when running under VirtualBox 1.6, but no luck. I installed the VirtualBox guest additions, but they didn't help. I followed the instructions in the Virtual Box manual to use the command line tool to define a custom video mode. Made no difference. I edited the xorg.conf file to add entries for 1280x1024 and 1440x1050. Nothing I do causes the Screen Resolution dialog to show me resolutions higher than 1024x768. I've googled for answers and followed all the tips I've found, but none of them are working for me. (Virtual Box is running under Windows XP on my laptop, if that matters.) Has anyone gotten this to work? My problem isn't exactly solved, but at least it's no longer an issue for me. See attached. ---BeginMessage--- Michael Thayer wrote: Hello Bill, Sending a size hint using VBoxManage while a VM is running (e.g. 'VBoxManage controlvm setvideomodehint 1280 1024 32') should work, and should override the limitation. Fullscreen mode and resizing the window using the mouse are also not affected. Ok, that's interesting... I don't know what effect the size hint had, if any. It didn't change the size of the window and it didn't cause the new size to appear in the screen resolution preferences. But... Apparently I haven't tried resizing the VirtualBox window recently, perhaps not since I added the guest additions. It used to be that when I resized the window or went to full screen mode, the OpenSolaris screen remained the same size and was centered in the VirtualBox window. Now the OpenSolaris screen changes size, including using the whole screen in full screen mode!!! Cool! I still don't know how to get additional choices in the screen resolution preferences, but with the ability to resize the screen that's not an important issue for me anymore. Thanks for your help!!! ---End Message--- ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] screen resolution problem with OpenSolaris 2008.05 on VirtualBox 1.6
Mike DeMarco wrote: Just a thought, How much video ram are you giving the guest? 32MB. I changed it to 128MB but it didn't make any difference. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] CDE libraries?
Are or will the CDE libraries be available for OpenSolaris? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] screen resolution problem with OpenSolaris 2008.05 on VirtualBox 1.6
I spent several hours trying to get OpenSolaris 2008.05 (final version) to work with a screen resolution larger than 1024x768 when running under VirtualBox 1.6, but no luck. I installed the VirtualBox guest additions, but they didn't help. I followed the instructions in the Virtual Box manual to use the command line tool to define a custom video mode. Made no difference. I edited the xorg.conf file to add entries for 1280x1024 and 1440x1050. Nothing I do causes the Screen Resolution dialog to show me resolutions higher than 1024x768. I've googled for answers and followed all the tips I've found, but none of them are working for me. (Virtual Box is running under Windows XP on my laptop, if that matters.) Has anyone gotten this to work? ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org