From tfb at tfeb.org Wed Oct 1 06:52:52 2008 From: tfb at tfeb.org (Tim Bradshaw) Date: Wed, 1 Oct 2008 11:52:52 +0100 Subject: Summary: how to register a Solaris subscription with SunSolve Message-ID: Thanks to Martin Pre_laber for his helpful response. I mailed / filled in various web forms, the one I think that was most useful was the one Martin recommended, which was this: http://sunsolve.sun.com/show.do?target=feedback . The result of this is that my sunsolve account now has a valid contract, so this process does seem to work. Except: it's not *my* contract/subscription, and its expiry is a long time before my subscription expires. So come December I will have to go through this all again, sigh. --tim -- Tim Bradshaw tfb at tfeb.org / +44 798 098 1475 PGP: 53AA CFA7 553B C611 B984 C4F2 0F11 EA99 0B7A 5ED1 _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From srinivas.banda at ge.com Wed Oct 1 13:23:49 2008 From: srinivas.banda at ge.com (Banda, Srinivas (GE, Corporate, consultant)) Date: Wed, 1 Oct 2008 13:23:49 -0400 Subject: SUMMARY: Server full backup and restore In-Reply-To: <184D5D31D664A445B4142B6CD9441ED90A1ECA49@STAMLVEM03.e2k.ad.ge.com> References: <184D5D31D664A445B4142B6CD9441ED90A1ECA49@STAMLVEM03.e2k.ad.ge.com> Message-ID: <184D5D31D664A445B4142B6CD9441ED90A1ECA4B@STAMLVEM03.e2k.ad.ge.com> Thanks to everyone for the replies and solutions. This is my first post in sunmanagers. Most of them suggested to use either flarcreate or ufsdump/ufsrestore. I'm going ahead with ufsdump/ufsrestore option. Thanks again... SB > _____________________________________________ > From: Banda, Srinivas (GE, Corporate, consultant) > Sent: Wednesday, October 01, 2008 9:31 AM > To: 'sunmanagers at sunmanagers.org' > Subject: Server full backup and restore > > Hi Sun Gurus > I need to take a Sun server full backup. There is no Veritas netbackup > installed and the server doesn't have any tape drive provision. Is it > possible to take the server's backup onto another server (say /dump > has sufficient free space) and restore it by nfs mounting that backup > mountpoint? Are there any other options? > > Thanks, > SB _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From Ryan.Anderson at baesystems.com Wed Oct 1 13:27:11 2008 From: Ryan.Anderson at baesystems.com (Anderson, Ryan C (US SSA)) Date: Wed, 1 Oct 2008 12:27:11 -0500 Subject: [SUMMARY] LDAP groups with many members + Sol 10 LDAP client Message-ID: <738m08$kcq47@dmzms99802.na.baesystems.com> The final answer is that you must patch upgrade Solaris 10 to see all members of large LDAP groups. I did find out the specific patch that fixes the behavior is the kernel patch. Solaris 10 SPARC kernel patch 120011-17 or above is supposed to fix this. BEWARE: I also found that the newest kernel patch (127127-11) makes printing from LDAP clients all but impossible, see: http://sunsolve.sun.com/search/document.do?assetkey=1-66-241426-1 RCA -- UNIX Administrator, BAE Systems EIT desk 763-572-6684 mobile 612-419-9362 -----Original Message----- From: sunmanagers-bounces at sunmanagers.org [mailto:sunmanagers-bounces at sunmanagers.org] On Behalf Of Anderson, Ryan C (US SSA) Sent: Tuesday, September 30, 2008 3:05 PM To: sunmanagers at sunmanagers.org Subject: LDAP groups with many members + Sol 10 LDAP client I've found that (like NIS) a Solaris 10 update 4 LDAP client can't see groups after they reach too many members. I found out today a Solaris 10 update 5 client can see the groups fine, but all my systems are update 4. The only workaround I've found is to create multiple groups with the same gidNumber and chunk up the members between them. Is there a setting in Solaris to see the groups properly? Any idea on what Sol 10 patch number might fix the behavior? On Sol 10 u4 & u5, I can do: ldaplist -l group verybiggroup But on u4, every other OS utility can't see the group, ie 'getent group verybiggroup', 'groups '. RCA -- UNIX Administrator, BAE Systems EIT desk 763-572-6684 mobile 612-419-9362 _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From Rob.McMahon at warwick.ac.uk Thu Oct 2 03:16:29 2008 From: Rob.McMahon at warwick.ac.uk (Rob McMahon) Date: Thu, 02 Oct 2008 08:16:29 +0100 Subject: SUMMARY: Sun Dual GigE on V890 In-Reply-To: <48D0D38D.7080607@warwick.ac.uk> References: <48CFE25A.9090205@warwick.ac.uk> <48D0D38D.7080607@warwick.ac.uk> Message-ID: <48E4754D.1050007@warwick.ac.uk> Rob McMahon wrote: >> I've just installed a shiny new X7285A Sun PCI-X Dual GigE UTP Low >> Profile card in a V890 running Solaris 10: >> >> SunOS narcissus 5.10 Generic_118833-36 sun4u sparc SUNW,Sun-Fire-V890 >> Solaris [... which the e1000g0 driver ignored ...] Having had the chance for some downtime, this was fixed by installing the latest 10_Recommended patch cluster. Cheers, Rob -- E-Mail: Rob.McMahon at warwick.ac.uk PHONE: +44 24 7652 3037 Rob McMahon, IT Services, Warwick University, Coventry, CV4 7AL, England _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From sunhux at gmail.com Fri Oct 3 07:18:43 2008 From: sunhux at gmail.com (sunhux G) Date: Fri, 3 Oct 2008 19:18:43 +0800 Subject: Summary: recovering a V120 with unknown eeprom password & inaccessible ALOM Message-ID: <60f08e700810030418n42815cf8je5c657ac0a0ce2c9@mail.gmail.com> Thanks to Rhys (D R Hughes), his reply appended below. I've also managed to recover from it : after waiting for 3 hrs, the "Console login:" prompt appeared & managed to login with root (lucky root lets me in). Then used "eeprom security-mode=none" to clear the eeprom password. Also, took the opportunity to trim down all those startup script (yes, Liebert's /etc/init.d/LiebertM script is the culprit that slowed down A LOM port). Reboot is now smooth & fast & without firmware password Thanks ========================================================= Have a look at Sun doc 816-2090-10.pdf (SunFire V120 and Netra 120 "Server User's Guide" from; http://docs.sun.com/app/docs/coll/netra120 ) >From memory (failing badly) with a serial terminal attached to the V120's LOM port (ie; first serial port) your type the LOM escape sequence #. (ie; "hash" and "dot" characters). Provided nobody has altered this default LOM escape sequence you should get the LOM> prompt. At the LOM> prompt you probably need to enter the equivalent LOM command to "Stop" plus "N" (or "L1" plus "N") poweron (ie; both keys must be held down as the machine is poweron and until the OK> prompt appears [so you probably can't do the equivalent from a serial term, but it may be worth trying]) key sequence to reset the NVRAM to its default settings. To do this on the V120, at the LOM> prompt enter "boot reset_nvram". Hopefully this will reset the V120's NVRAM to factory defaults. However the unknown password will probably still be stored in memory even though its been disabled. So its a good idea to change the passwd to one you know and store safe, enable it to test it and then decide whether or not to keep it enabled or to disable it. I think that the NVRAM settings are stored on the "System Configuration Card", so if you have another compatible SunFire machine you could try changing the card (ie; whilst they're powered off). On Fri, Oct 3, 2008 at 4:58 PM, sunhux G wrote: > Hi > > I've inherited a V120 with unknown eeprom password which I intend > to reinstall with Solaris 10 & then use it as NFS/Samba server. > > When this V120 boots up, it would hang the console (ie my notebook's > Hyperterm session) while it's firing up Liebert's Multilink script. > > Stop-A (Ctrl-Break in Hyperterm which only works occasionally) > brings up this message : > Type go(continue), boot or login. > Any of the above 3 options will prompt for a (eeprom) password > > I happen to be lucky at one point that I got a "Console login:" > prompt after waiting for 2 hours watching the LIebert Multilink UPS > messages at the console/hyperterm. I was dumb not to have > seized the earlier opportunity to use "eeprom security-password=" > when I managed to get the chance earlier to login but instead > shut it down to try to get to OBP (ok prompt) which did not > materialize as I later found it needs an eeprom password. > > My notebook has a serial cable connected to the ALOM port > on this V120 - gets to see the bootup messages till the point > Liebert Multilink messages show up. > > Any way for me to recover from this so that I could boot from > Solaris 10 DVD & reinstall? Does V120 has a little processor > (like V880) that requires us to press certain key combination/ > sequence to get to OBP/ok prompt mode? > > > > Thanks > U _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From fabiomiranda at ice.co.cr Sat Oct 4 01:17:30 2008 From: fabiomiranda at ice.co.cr (Fabio A. Miranda) Date: Fri, 03 Oct 2008 22:17:30 -0700 Subject: Summary: Boot disk mirror using SVM Message-ID: <48E6FC6A.4010907@ice.co.cr> Thanks for all the answers ( metadevices is a hot topic in Solaris gurus fields) Problem: Fresh install of Solaris 10 on V440 (2 x 72GB FC-AL). Slice for metadb was not present, not free space, while using format came across "partition mounted" type of errors (even in single user mode). Solution: I. V440 has raid controler, it's straightforward to use it via: raidctl ( man option -c). Note that HW mirroring is MUCH better than Software one. Always count on HW mirroring if present. II. Got one answer that claims Solaris 10 made a bit change that requires, even in single usermode with partition unmounted, to exclude partition intented to be reduced out of /etc/vstab. There is also a shell variable you set via export ( sorry, I delete the e-mail ). So, let's say I want to get some extra space out of /export, I go to single usermount, umount /export, vi /etc/fstab, comment /export entry, format, par, p, and then add new slice ( FYI: s7 is usually used to metadb by Sun practices). Also, remmber to metadb -a -f -c 3 /dev/dsk/cXtXdXsX III. ufsdump/ufsrestore the current partitions to c1t1 and boot from there, then repartition c1t0. Thanks again, as usual, sunmanagers is a true source of expertise and knowledge. fabio. _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From cbarnar1 at earthlink.net Sun Oct 5 06:05:11 2008 From: cbarnar1 at earthlink.net (Christopher L. Barnard) Date: Sun, 5 Oct 2008 05:05:11 -0500 Subject: Summary: jumpstart and Packages In-Reply-To: <4FB1EFF3-1707-481B-A3DA-DE571FB3C587@earthlink.net> References: <4FB1EFF3-1707-481B-A3DA-DE571FB3C587@earthlink.net> Message-ID: <58692A33-37C4-4135-9562-BEDFA443CE76@earthlink.net> I asked > In our environment, /usr/local and /opt/local are the same. Usually > this > is accomplished by having /usr/local be a symlink to /opt/local. But > occasionally there are old systems that have /opt/local as a symlink > back > to /usr/local either because the local stuff won't fit in /opt but > will > fit in /usr or because /usr/local is a separate filesystem that just > happens to be mounted as /usr/local. > > I have a package that installs into /[usr|opt]/local/bin just fine > when installed interactively. It has no problem with the symlink. > > pkgadd -d OpenSSH*. > > But when I put it into jumpstart, the jumpstart barfs on the symlink: > > [...] > pkgadd: ERROR: unable to create package object . > pathname does not exist > unable to create directory > /a/usr/local/bin > /a/usr/local/bin/scp > pkgadd: ERROR: unable to create directory : (17) File > exists > [...] > > (in jumpstart, the finished server image is mounted as /a). > > Note that it is trying to create the directory /usr/local. Yes it > exists. > Its a symlink. My jumpstart rules created it about five lines earlier > than this package install. > > Does anyone have any idea why > (1) this works interactively but not in jumpstart, and > (2) what I need to do to get this to work? I build these packages > myself, > so I am quite comfortable editing the pkg image build if that is > needed. > I intentionally do not have in the package prototype the definition > of /usr/local so that both a symlink and an existing directory work. > The package installation just acts with the blind trust that > /usr/local exists in some form or other. The answer: yes, /usr/local exists. It is a symlink to /opt/local which is not the same as /a/opt/local. I changed it to a relative symlink (/usr/ local -> ../opt/local) and it works like a charm. Thanks to the many many folks who pointed out the bad symlink. Christopher L. Barnard cbarnar1 at earthlink.net ----------------------------------------------------------------------- When I was a boy, I was told that anyone could be president. Now I am beginning to believe it. -- Clarence Darrow _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From Crist.Clark at globalstar.com Tue Oct 7 15:16:37 2008 From: Crist.Clark at globalstar.com (Crist Clark) Date: Tue, 07 Oct 2008 12:16:37 -0700 Subject: SUMMARY: Setting a 255.255.255.255 Netmask In-Reply-To: <48E9F2F7.33E4.0097.0@globalstar.com> References: <48E9F2F7.33E4.0097.0@globalstar.com> Message-ID: <48EB5320.33E4.0097.0@globalstar.com> I got a number of replies, mostly people wondering why the heck I was doing this. Believe me, I do have a reason. But thanks to Peter van Gemert who helpfully pointed out that the following syntax, zonecfg:zone1> add net zonecfg:zone1:net> set physical=bge0 zonecfg:zone1:net> set address=192.168.222.18/32 zonecfg:zone1:net> end Is valid and works as desired. I went back to the documentation, zonecfg(1M) page, and sure enough, this is in there, but I never noticed it before, net: address, physical The network address and physical interface name of the network interface. The network address is one of: o a valid IPv4 address, optionally followed by "/" and a prefix length; But the consensus seems to be that the netmasks(4) file is a little too smart for its own good and the libraries backing it don't deal with >30-bit netmasks well. On 10/6/2008 at 11:14 AM, "Crist Clark" wrote: > I've got a situation where I want to assign a 255.255.255.255 > netmask to an interface. It's easy enough to do with the > ifconfig(1M) command, > > # ifconfig bge2 plumb 192.168.222.18 netmask 255.255.255.255 broadcast > 255.255.255.255 > # ifconfig bge2 > bge2: flags=1000802 mtu 1500 index 6 > inet 192.168.222.18 netmask ffffffff broadcast 255.255.255.255 > ether 0:14:4f:a2:ee:98 > > But if I add, > > 192.168.222.18 255.255.255.255 > > To the netmasks(4) file, I get this, > > # ifconfig bge2 plumb 192.168.222.18 netmask + broadcast + > Need net number for mask > # ifconfig bge2 > bge2: flags=1000802 mtu 1500 index 7 > inet 192.168.222.18 netmask ffffff00 broadcast 192.168.222.255 > ether 0:14:4f:a2:ee:98 > > It gives an error message and the classful (in this case, Class C0) > address for the network number. > > Now, I could just do this "manually" in the hostname.bge2 file, but > I actually want to do this in a non-global zone. I don't know how > to pass extra ifconfig(1M) parameters into the configuration of a > zone's interfaces via the "add net" in zonecfg(1M). How do I get > a 255.255.255.255 netmask on a non-global zone. I guess I could still > kludge up a hostname.bge2:1 to do it, but is there a "cleaner" way? > (I know of whole 'nother approach to this, but have never tried that > in a zone either.) _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From fabiomiranda at ice.co.cr Wed Oct 8 13:28:45 2008 From: fabiomiranda at ice.co.cr (Fabio A. Miranda) Date: Wed, 08 Oct 2008 11:28:45 -0600 Subject: Summary: IP kern.notice message dst FFFFFFFF src 7F000001 Message-ID: <48ECEDCD.3020800@ice.co.cr> Thanks for all the answers: Given on a m4000: Oct 3 18:11:30 cltrde3 ip: [ID 390400 kern.notice] dst FFFFFFFF src 7F000001 This messages means that a device on the network is using 127.0.0.1 as its source address. The message can be ignored. If you have time, snoop interface to figure out source MAC and "catch" the (faulty?) device. Another solution mentioned was redirect kern.notice to /dev/null. It is up to each sys-admin wether this can be accomplish because this will apply to all applications using that facility. Thanks again, fabio _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From jesse-carroll at usa.net Thu Oct 9 17:24:32 2008 From: jesse-carroll at usa.net (JESSE CARROLL) Date: Thu, 09 Oct 2008 17:24:32 -0400 Subject: Summary: Hardware mirroring on t5220 without reboot Message-ID: <737miyspZ7122S14.1223587472@cmsweb03.cms.usa.net> Sorry about the delay, but vacations are nice.... The answer from Sun is no, you have to reboot the system. It has to do with the hardware raid controller not being the brightest bulb in the pack. In fact I discovered that the controller ONLY allows 2 raid volumes. Given that a T5220 can have up to 8 drives means that you have to use something else (SVM or Veritas) to create more protected volumes. >From the responses I received I guess I wasn't clear about what I want to do: 1) I can see the new "disk" using raidctl -l and raidctl -l c1t2d0. 2) I did wait until the mirroring completed before trying anything else 3) I can not see the new "disk" using format, without a boot -r. 4) For the first mirror I booted off the network. 5) I'm trying to use the second raid volume without rebooting the system. ------ Original Message ------ Received: Thu, 25 Sep 2008 08:59:57 AM EDT From: "JESSE CARROLL" To: Subject: Hardware mirroring on t5220 without reboot After hardware mirroring the first set of drive on a T5220 and installing the OS on this mirror (c1t0d0) I want to mirror some of the remaining 6 drives. The second mirror completes but I don't see the device (i.e. c1t2d0) with format or in the device tree. I do see it with raidctl. The manual doesn't mention this senario. I've tried 'devfsadm -C' and 'cfgadm -c configure c1::ds/c1t2d0" to no avail. I'm running Sol 10 0508 with the current patch bundle. Is there a way to "see" the new raid device without a reboot? JC _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From joe_fletcher at btconnect.com Wed Oct 15 10:21:05 2008 From: joe_fletcher at btconnect.com (joe fletcher) Date: Wed, 15 Oct 2008 15:21:05 +0100 Subject: SUMMARY (partial): UFS slowness on Opteron References: <7FC614F7CE1FCE449C3674BAD9A52032AC74BB@HEMV3BUKER.he.local> Message-ID: <7FC614F7CE1FCE449C3674BAD9A52032AC74C3@HEMV3BUKER.he.local> It's memory management on the Opterons. If we rip out most of the DIMMs it speeds up. The machine originally had 256Gb and was pushing data half the speed of our other machines with 128Gb fitted. If we run the tests with 32Gb it behaves fine. We're playing with the hardware config to find out where the tiping point is. Cheers Joe ________________________________ Hi, This is a bit woolly so apologies in advance. Got some new x4600, 8x quad core Opteron boxes, 256Gb RAM attached to EMC Clariion cx380s via Emulex LP10K HBAs. Kernel rev is 127112-11. We have a big financial app that we're trying to migrate from a 6900 but when it comes to filesystem performance it's dragging. We've done some testing with iozone and the Opterons seem to be consistently slower than the SPARCs. For example on one of our write tests the AMDs do it in approx 1 min whereas an old v490 can do the same test in under 20s. We discovered last night that if we use ZFS we can get the AMDs to behave however we're not sure about ZFS support for the application. We've tried various tuning eg maxphys, sd-max-transfer, sd throttling and so on. Things that seem to work on the SPARCs provide no apparent benefit on the AMDs even though the docs indicate they should. Does anyone know of where we might look to get UFS to perform up to expectation? Cheers Joe _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From cbarnar1 at earthlink.net Wed Oct 15 19:49:47 2008 From: cbarnar1 at earthlink.net (Christopher L. Barnard) Date: Wed, 15 Oct 2008 18:49:47 -0500 Subject: Summary: Process being killed. In-Reply-To: <8D76EE12-7262-4B01-B678-B66D289CCCFF@earthlink.net> References: <8D76EE12-7262-4B01-B678-B66D289CCCFF@earthlink.net> Message-ID: I asked: > This is Solaris 10 and I have a sneaky feeling that may be the > problem. > > I formerly worked with commercial Tripwire at a previous installation > and I am working on setting up the OpenSource Tripwire with my current > company. I have customized the policy file as I want. When I run it > from the command line, this works just as I want it to. However, when > I run it from cron it is summarily killed. > > Your "cron" job on sandbox > /opt/local/tripwire/sbin/tripwire --check --email-report > /dev/null > 2>&1 > > produced the following output: > > Killed > > I expect it has to do with the difference between environments in cron > and in an interactive shell. All binaries and scripts in the policy > and the config are fully pathed, so it is not a path issue. I can run > it from the command line in /bin/sh, so it is not a shell issue. I am > left wondering if it is a resource issue, except that that was never > an issue with commercial TFS on Solaris 10 servers. can anyone > suggest what else to try? The solution: an ldd of the tripwire binary showed that libraries in both /usr/lib and /usr/local/lib were being referenced. So I changed the cron entry to explicitly set the LD_LIBRARY_PATH: 03 04 * * * LD_LIBRARY_PATH=/usr/lib:/usr/local/lib; export LD_LIBRARY_PATH; /opt/local/tripwire/sbin/tripwire --check --email- report > /dev/null 2>&1 All is well. Thanks for all of the suggestions. I had forgotten to include in my initial question that I had already redirected to a file rather than to /dev/null and learned nothing. Thanks to cmorris at cs.odu.edu ric at Opus1.COM luc at suryo.com maciej.blizinski at gmail.com john.stoffel at taec.toshiba.com mid.xinef at gmail.com ddunham at taos.com alanpae at ilkda.com Andrew.Henle at shomo.com Christopher L. Barnard cbarnar1 at earthlink.net ----------------------------------------------------------------------- When I was a boy, I was told that anyone could be president. Now I am beginning to believe it. -- Clarence Darrow _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From cbarnar1 at earthlink.net Wed Oct 15 20:01:18 2008 From: cbarnar1 at earthlink.net (Christopher L. Barnard) Date: Wed, 15 Oct 2008 19:01:18 -0500 Subject: Summary: gzip and logadm not playing nicely together In-Reply-To: <2D614F13-9BDD-41B0-95AE-9E41DFE7C0EC@earthlink.net> References: <2D614F13-9BDD-41B0-95AE-9E41DFE7C0EC@earthlink.net> Message-ID: I asked: > Solaris 9 server. > > logadm.conf has line > /var/log/syslog -C 8 -P 'Thu Oct 9 08:10:00 2008' -a 'kill -HUP `cat > /var/run/syslog.pid`' -c -p 1d -z 2 > > Note that these files are very large (on the order of 1gig in size > after > 24 hours) so that is why they are rotated daily. > > syslog rotates to syslog.0 to syslog.1 just fine. syslog.2.gz rotates > to > syslog.3.gz to syslog4.gz to syslog.5.gz to syslog.6.gz just fine. > > The server seems to lose its mind when it comes to rotating syslog.1 > to > syslog.2 and then compressing it. I get every night: > > Date: Thu, 9 Oct 2008 03:10:18 -0500 (CDT) > From: root on rushmail > To: root at rushmail.cc1.rpslmc.edu > Subject: Output from "cron" command > > Your "cron" job on rushmail > /usr/sbin/logadm > > produced the following output: > > logadm: Warning: command failed: /bin/gzip /var/log/syslog.2 > gzip: /var/log/syslog.2.gz already exists; not overwritten > > When I log in I see that a zero-size syslog.2.gz exists, and .1 has > been > rotated to .2. So I manually delete the zero-size syslog.2.gz and > compress syslog.2 every morning. Does anyone know of what trick is > needed > to get this to work automatically? The short answer: patch 116016-06. The long answer: the fact that -c and -z do not initially behave themselves together is documented in http://sunsolve.sun.com/search/document.do?assetkey=1-1-4885799-1 Unfortunately that did not do the trick. I opened a Sun Case ID and they suggested I apply this patch. This patch actually implements this workaround, but also several other bug fixes. Thanks to: ghicks at cadence.com mh1272 at gmail.com florkle at yahoo.com (Who gave me the sunsolve link) Crist.Clark at globalstar.com DRoss-Smith at reviewjournal.com hvjunk at gmail.com Christopher L. Barnard cbarnar1 at earthlink.net ----------------------------------------------------------------------- When I was a boy, I was told that anyone could be president. Now I am beginning to believe it. -- Clarence Darrow _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From bryan.bahnmiller at managedmail.com Wed Oct 15 10:55:21 2008 From: bryan.bahnmiller at managedmail.com (Bryan Bahnmiller) Date: Wed, 15 Oct 2008 09:55:21 -0500 Subject: SUMMARY (partial): UFS slowness on Opteron In-Reply-To: <7FC614F7CE1FCE449C3674BAD9A52032AC74C3@HEMV3BUKER.he.local> References: <7FC614F7CE1FCE449C3674BAD9A52032AC74BB@HEMV3BUKER.he.local> <7FC614F7CE1FCE449C3674BAD9A52032AC74C3@HEMV3BUKER.he.local> Message-ID: <48F60459.80308@managedmail.com> We were trying to get performance out of our x4600's too. I found a x4600 manual on Sunsolve that explained things. "With just two or four DIMMs populated, the system runs memory at 667MHz. With 6 or 8 DIMMs populated, the system runs memory at 533 MHz." You must fill out the DIMMs from top to bottom. If you fill up only the white DIMM slots you will be running at maximum memory speed. Once you use the black DIMM slots your memory speed drops by 20%. So if you ordered the system with 16 GB RAM per CPU module, Sun will typically configure it as 2GBx8. We had to change our order to get 4GBx4. Bryan joe fletcher wrote: > It's memory management on the Opterons. If we rip out most of the DIMMs it > speeds up. The machine originally had 256Gb and was pushing data half the > speed of our other machines with 128Gb fitted. If we run the tests with 32Gb > it behaves fine. We're playing with the hardware config to find out where the > tiping point is. > > Cheers > > Joe > > ________________________________ > > > > Hi, > > This is a bit woolly so apologies in advance. > > Got some new x4600, 8x quad core Opteron boxes, 256Gb RAM attached to EMC > Clariion cx380s via Emulex LP10K HBAs. Kernel rev is 127112-11. We have a big > financial app that we're trying to migrate from a 6900 but when it comes to > filesystem performance it's dragging. > > We've done some testing with iozone and the Opterons seem to be consistently > slower than the SPARCs. For example on one of our write tests the AMDs do it > in approx 1 min whereas an old v490 can do the same test in under 20s. > > We discovered last night that if we use ZFS we can get the AMDs to behave > however we're not sure about ZFS support for the application. > > We've tried various tuning eg maxphys, sd-max-transfer, sd throttling and so > on. Things that seem to work on the SPARCs provide no apparent benefit on the > AMDs even though the docs indicate they should. > > Does anyone know of where we might look to get UFS to perform up to > expectation? > > Cheers > > Joe > _______________________________________________ > sunmanagers mailing list > sunmanagers at sunmanagers.org > http://www.sunmanagers.org/mailman/listinfo/sunmanagers > _______________________________________________ > sunmanagers mailing list > sunmanagers at sunmanagers.org > http://www.sunmanagers.org/mailman/listinfo/sunmanagers _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From dave.markham at fjserv.net Fri Oct 17 07:18:40 2008 From: dave.markham at fjserv.net (Dave Markham) Date: Fri, 17 Oct 2008 12:18:40 +0100 Subject: SUMMARY: ssh issue within script Message-ID: <48F87490.3040102@fjserv.net> After some messing around (trial and error really) i have got it to work. Original problem is below but the resolution was to remove the -n from the ssh_opts variable and place it only on the source side of the ssh command. I had tried various -t -t options as i understand this relates to pseudo terminals but had no joy. The following now works :- ------ ssh_opts="-o ConnectTimeout=15 \ -o BatchMode=yes \ -o StrictHostKeyChecking=no \ -o LogLevel=quiet \ " #-n " ----------- ssh ${ssh_opts} -n 2>> ${err_log} user@${srv} "cat ${source_file}" | ssh ${ssh_opts} ${dest_host} "cat > ${dest_dir}/${srv}.latest.tar" 2>> ${err_log} The -n option prevents reading from standard in and allows ssh to run in the background and not override the shell script process. This is why i assume it is required on the first ssh and needs not to be present on the second. Hope this helps someone. Original request :- ---------------------------------- Guys i'm trying to run an ssh command from within a shell script but am getting an error. It works fine from the shell :- ------ ssh_opts="-o ConnectTimeout=15 \ -o BatchMode=yes \ -o StrictHostKeyChecking=no \ -o LogLevel=quiet \ -n " ----------- ssh ${ssh_opts} 2>> ${err_log} user@${srv} "cat ${source_file}" | ssh ${ssh_opts} "cat > ${dest_dir}/${srv}.latest.tar" 2>> ${err_log} The error i'm getting in my error log is this :- Pseudo-terminal will not be allocated because stdin is not a terminal. ----------------------------- The reason i am doing this type of transfer is i need to get files from one machine to another using a third as a middle man without storing the file on the middle machine. e.g host B initiates an ssh to host A and cats the file through its connection to a third host C. Any help appreciated Cheers _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From Sugan.Moodley at standardbank.co.za Sun Oct 19 07:14:10 2008 From: Sugan.Moodley at standardbank.co.za (Moodley, Sugan S) Date: Sun, 19 Oct 2008 13:14:10 +0200 Subject: SUMMARY: Urgent NFS mount problem - NFS fsstat failed for server Message-ID: <6BA6A13C6795174BBADC8AF67C07919002953C0B@00172MSGJNB1012.za.sbicdirectory.com> Hi all, I had a problem similar to one posted below by Alan but could not find a full summary. I managed to fix my problem so here's the initial problem and then the summary: NFS Server A shares a writable share to a specific client B however client B cannot write to it. Both servers running Solaris 10 sparc Observations and checks: - /etc/dfs/dfstab on Server A is: /usr/sbin/share -F nfs -o nosuid,rw=clientB /zones/clientA/FileSystems/usr/sap/trans - all nfs services on Server A running otherwise run: svcadm -v enable -r nfs/server - autofs is running on Client B and is accessing share via /net/ServerA/zones/clientA/FileSystems/usr/sap/trans - ensure that files and dir's on Server A has the same UID/GID numeric settings as on Client B - ensure all permissions in the full dir path of the share on Server A are at least 555 (read+execute) - we are using DNS in /etc/resolv.conf and /etc/nsswitch.conf SUMMARY: Change the share to point the fully qualified domain name of Client B and the access of Server A from Client B must also be the same: - /etc/dfs/dfstab on Server A: /usr/sbin/share -F nfs -o nosuid,rw=clientB.domain.name.com /zones/clientA/FileSystems/usr/sap/trans - unshare /zones/clientA/FileSystems/usr/sap/trans - shareall - on Client B: cd /net/ServerA.domain.name.com/zones/clientA/FileSystems/usr/sap/trans - touch a This was not a problem until we install some Live Upgrade minimum patches recently (refer to latest version of SUN's infodoc 72099) Regards Sugan Moodley Original SUMMARY: PARTIAL SUMMARY: Urgent NFS mount problem - NFS fsstat failed for server Alan Bradley - CPX WC AlanB at ComparexAfrica.co.za Mon Jan 27 10:39:19 EST 2003 ________________________________ Hi All, By sharing the filesystems as 'public' I could get them mounted. However, if I share specifically for machine 'abcd' (-o rw=abcd) I get the errors as below. Regards, Alan. -----Original Message----- From: Alan Bradley - CPX WC [mailto:AlanB at comparexafrica.co.za ] Sent: Monday, January 27, 2003 4:38 PM To: Sun Managers List Subject: Urgent NFS mount problem - NFS fsstat failed for server Hi All, I have an urgent problem with some NFS mounts: When I try to mount the NFS directories, I get a permission denied error. If I do a df -k I get the following output: abcd:root:/etc/rc2.d => df -k Filesystem kbytes used avail capacity Mounted on /dev/md/dsk/d0 8261393 7341688 837092 90% / /proc 0 0 0 0% /proc fd 0 0 0 0% /dev/fd mnttab 0 0 0 0% /etc/mnttab swap 4837216 16 4837200 1% /var/run swap 4860336 23136 4837200 1% /tmp /dev/md/dsk/d3 4131866 2486684 1603864 61% /backup /dev/md/dsk/d4 216325417 192114915 22047248 90% /db2/dms NFS fsstat failed for server wxyz: error 7 (RPC: Authentication error) df: cannot statvfs /opt/SUNWspro: Permission denied abcd:root:/etc/rc2.d => The following appears in /var/adm/messages: Jan 27 16:24:07 abcd last message repeated 109 times Jan 27 16:24:38 abcd nfs: [ID 664466 kern.notice] NFS getattr failed for server wxyz: error 7 (RPC: Authentication error) I have checked/done the following: 1. Restarted the NFS daemons on server: wxyz 2. Restarted the NFS daemons on server: abcd 3. Changed the mount point permissions to '777' 4. If I run a showmount -e OR a dfshares on server wxyz the directories are exported 5. If on server abcd, and I run: dfshares wxyz I also see the directories available I am now out of idea's and urgently need some help please! Unfortunately I cannot reboot the servers at the moment. I will summarise. Regards, Alan. _____________________________________________________________________________ ________________________________________________________ Standard Bank email Disclaimer and confidentiality note This e-mail, its attachments and any rights attaching hereto are, unless the content clearly indicates otherwise, the property of Standard Bank Group Limited and its subsidiaries. It is confidential, private and intended for only the addressee. Should you not be the addressee and receive this e-mail by mistake, kindly notify the sender, and delete this e-mail immediately. Do not disclose or use it in any way. Views and opinions expressed in this e-mail are those of the sender unless clearly stated as those of Standard Bank Group. Standard Bank Group accepts no liability for any loss or damages howsoever incurred, or suffered, resulting, or arising, from the use of this email or its attachments. Standard Bank Group does not warrant the integrity of this e-mail nor that it is free of errors, viruses, interception or interference. Licensed divisions of the Standard Bank Group are authorised financial services providers in terms of the Financial Advisory and Intermediary Services Act, No 37 of 2002 (FAIS). For information about the Standard Bank Group visit our website http://www.standardbank.com _____________________________________________________________________________ ___________________________________________________________ From hnulty at email.unc.edu Tue Oct 21 11:25:22 2008 From: hnulty at email.unc.edu (Helen Nulty) Date: Tue, 21 Oct 2008 11:25:22 -0400 Subject: Summary: replace primary disk in SunFire V100 Message-ID: <48FDF462.9030107@email.unc.edu> We appreciate all the response. Consensus is: the disk from the SunBlade 150 WILL work, that any regular ata/ide disk will work; disk replacement will be a standard procedure. Our system was mirrored across primary and secondary disks; the secondary disk is carrying the load now. We are not facing a re-install - just need to remirror with disksuite once the bad disk is replaced. The initial problem was the primary disk lost its label. We got "corrupt label - bad geometry" on a routine reboot. We were able to boot from the secondary disk. (phew!) We ran format to repair the label and found the primary disk no longer matched the secondary? Very curious...or is this to be expected? It had 65529 cylinders compared to 19156 on the secondary; 31.5 GB compared to 37.27 GB on the secondary. We were able to partition it and label it using the new smaller size; fsck of raw partitions failed until a filesystem was built on them - then fsck succeeded. We are unable to duplicate the original partition table. This is what happens: Warning: error writing VTOC. Oct 21 08:27:45 soda dada: WARNING: /pci at 1f,0/ide at d/dad at 0,0 (dad1): Oct 21 08:27:45 soda ID not found Oct 21 08:27:46 soda dada: WARNING: /pci at 1f,0/ide at d/dad at 0,0 (dad1): Oct 21 08:27:46 soda disk not responding to selection DIOCTL_RWCMD: I/O error Warning: error reading backup label. Have you ever seen anything like this? Is this disk salvageable? Thanks for your help...will summarize for the list again with any new information. Helen -- Helen P. Nulty Systems Analyst UNC-CH School of Public Health Instructional & Information Systems CB# 7400 Chapel Hill, NC 27599 ph: 919-966-9136 email: Helen_Nulty at unc.edu _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From dani.calloway at gmail.com Tue Oct 21 21:41:15 2008 From: dani.calloway at gmail.com (Dani Calloway) Date: Tue, 21 Oct 2008 17:41:15 -0800 Subject: SUMMARY: jumpstarting and ILOM Message-ID: Often mentioned was the remote access capabilities of the ILOM. While they sound interesting, we have an existing serial-console-management solution, and don't want to deviate from it for a few systems. To that end... Rather than waiting for the right moment in the boot cycle and hitting break, you can instruct the ILOM to not boot automatically. To do this, run 'set /HOST/bootmode script="setenv auto-boot? false"', or "set auto-boot?=false" if you're already in OBP. Then, when the system is powercycled, you'll always get an OK prompt before continuing. This allows you to netboot or boot with non-standard arguments. Thanks to Pedro Baldanta, Maciej BliziEski, Matt Clausen, and Andrew Caines for their responses to this effect. -- Dani Calloway On 10/3/08, Dani Calloway wrote: > We recently received some new Sun SPARC T5120 servers, and we need to > jumpstart them. However, they came with a new "feature" which Sun > calls ILOM - integrated lights out manager. With the ILOM, however, I > can't figure out how to instruct the system to boot (and, thus, > install) via the network. I have to reboot the system, hit Break at > the exact correct moment during the boot to get an OK prompt and > execute "boot net - install" from there. > > So, my question is: is there a way to make the system boot from > network while in the ILOM shell, or do I have to continually reboot > and break the boot sequence until I get the timing right and can > netboot from the OK prompt? > > -- > Dani Calloway _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From sanelson at gmail.com Wed Oct 22 11:22:34 2008 From: sanelson at gmail.com (Stephen Nelson-Smith) Date: Wed, 22 Oct 2008 16:22:34 +0100 Subject: SUMMARY: Hardware for Oracle Serverq Message-ID: Hello all, Apologies for the late summary - I've only just got around to acting on your many recommendations. Here are the salient points: * Many of you pointed out that at my price point, the x86 systems are better performers and better value for money. * A number of you recommended I run Oracle on Linux at my price point However, my vendor tells me Linux is not supported, nor is Solaris on x86, so that pushed me to SPARC. * It was observed that since the discontinuation of the v240/245 series, there's a bit of a hole in the range * Getting a refurbished or second-hand SPARC IV box was a strong recommendation, which I shared, but my FD wanted to buy new * There were some reservations around running Oracle on the Niagra chipset, particularly around how well the thread-sharing model works with Oracle, and whether the FP model would be a problem. I did some research myself, and found a number of people running OLTP applications on Niagra chips, very happily. * The T1000 was out of the question - not enough disks * The T2000 looked like an option, as did the 5120 I've opted for the 5120 and stretched the budget. Thanks for your many thoughts. S. _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From bj+sunmanagers at glue.ch Wed Oct 22 14:08:59 2008 From: bj+sunmanagers at glue.ch (Beat Jucker) Date: Wed, 22 Oct 2008 20:08:59 +0200 Subject: Summary: SSH X11 forwarding: No xauth data Message-ID: <20081022200859.A9444@sol-dev1.glue.ch> Thanks to Udo Grabowski and Ric Anderson for their fast replies. I only compared the system configuration files in /etc/ssh but not the users config file ~user/.ssh/config. # grep XAuthLocation ~zuser/.ssh/config XAuthLocation /usr/openwin/bin # ls -l ~zuser/.ssh/config -rw-r--r-- 1 zuser glue 112 Mar 17 2000 /home/zuser/.ssh/config Deleting XAuthLocation entry solved the problem. Because this configuration was seen by different users on our side just on this particular system it could be possible that years ago we have used openssh and this configuration defined just the directory. Anyway, now it is working as expected with the solaris version of ssh. Thanks again! -- Beat _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From joe_fletcher at btconnect.com Mon Oct 27 07:51:03 2008 From: joe_fletcher at btconnect.com (joe fletcher) Date: Mon, 27 Oct 2008 12:51:03 -0000 Subject: SUMMARY (update): UFS slowness on Opteron References: <7FC614F7CE1FCE449C3674BAD9A52032AC74BB@HEMV3BUKER.he.local> <7FC614F7CE1FCE449C3674BAD9A52032AC74C3@HEMV3BUKER.he.local> Message-ID: <7FC614F7CE1FCE449C3674BAD9A52032AC74D5@HEMV3BUKER.he.local> Hi All, SUN have come up with a more permanent fix which is the addition of a virtual page manager (VPM) to the solaris kernel. It was due to become part of the next release of Solaris anyway but we've been given an advanced update to implement it. Things are now behaving rather more as expected given the hardware's raw capability. For those that care you can read around the subject if you look for BugID 6256083 and related docs. Cheers Joe ________________________________ From: joe fletcher Sent: Wed 15/10/2008 15:21 To: sunmanagers at sunmanagers.org Subject: SUMMARY (partial): UFS slowness on Opteron It's memory management on the Opterons. If we rip out most of the DIMMs it speeds up. The machine originally had 256Gb and was pushing data half the speed of our other machines with 128Gb fitted. If we run the tests with 32Gb it behaves fine. We're playing with the hardware config to find out where the tiping point is. Cheers Joe ________________________________ Hi, This is a bit woolly so apologies in advance. Got some new x4600, 8x quad core Opteron boxes, 256Gb RAM attached to EMC Clariion cx380s via Emulex LP10K HBAs. Kernel rev is 127112-11. We have a big financial app that we're trying to migrate from a 6900 but when it comes to filesystem performance it's dragging. We've done some testing with iozone and the Opterons seem to be consistently slower than the SPARCs. For example on one of our write tests the AMDs do it in approx 1 min whereas an old v490 can do the same test in under 20s. We discovered last night that if we use ZFS we can get the AMDs to behave however we're not sure about ZFS support for the application. We've tried various tuning eg maxphys, sd-max-transfer, sd throttling and so on. Things that seem to work on the SPARCs provide no apparent benefit on the AMDs even though the docs indicate they should. Does anyone know of where we might look to get UFS to perform up to expectation? Cheers Joe _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From ghc4+ at pitt.edu Mon Oct 27 14:18:27 2008 From: ghc4+ at pitt.edu (George Cebulka) Date: Mon, 27 Oct 2008 15:18:27 -0400 Subject: Summary - ILOM break sequence for an x4200???? Message-ID: <49061403.5010304@pitt.edu> Thanks everyone for your replies. The escape/break sequence to get from the OS to the SP is [Esc] + [Shift 9] or Hit the Escape key followed by a ( (Shift+9) Just in case, to get from the SP to the OS, start /SP/console - George Cebulka Senior System Engineer University of Pittsburgh Network Operations Center ghc4 at pitt.edu _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From maurice.akkermans at tellabs.com Wed Oct 29 02:50:12 2008 From: maurice.akkermans at tellabs.com (Akkermans, Maurice) Date: Wed, 29 Oct 2008 09:50:12 +0200 Subject: Summary: boot different server from disk In-Reply-To: <8E3E1A003EE07E409D0AF5D91EEBFD6A06946859@FIESEX1.tellabs-east.tellabsinc.net> References: <8E3E1A003EE07E409D0AF5D91EEBFD6A06946859@FIESEX1.tellabs-east.tellabsinc.net> Message-ID: <8E3E1A003EE07E409D0AF5D91EEBFD6A069BAB34@FIESEX1.tellabs-east.tellabsinc.net> Thank you all for the input. There are some ways to do this, but as the servers might be different makes the process more complex, with booting from alternative media, CD or net. As it is not a Sun supported solution we are going to look at a Cluster solution or maybe using LDOMs. Regards, Maurice -----Original Message----- From: sunmanagers-bounces at sunmanagers.org [mailto:sunmanagers-bounces at sunmanagers.org] On Behalf Of Akkermans, Maurice Sent: Thursday, October 23, 2008 14:03 To: sunmanagers at sunmanagers.org Subject: boot different server from disk I have a T2000 server (call it A) booting from a SAN disk. I want to boot another server (B) from this disk, after server A is shutdown. This server B presents the device at a different device path. How can I get around this? I have tried to boot server B in single user mode to make changes to /etc/path_to_inst and the /dev links, but the root disk seems to be read-only, while the mount command shows it is read-write: / on /pci at 7c0/pci at 0/pci at 8/SUNW,qlc at 0/fp at 0,0/disk at w210000d0231e6fe6,0:a read/write/setuid/devices/dev=1d80000 on Thu Jan 1 01:00:00 1970 Can you boot a server on another servers disk? Regards, Maurice ============================================================ The information contained in this message may be privileged and confidential and protected from disclosure. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any reproduction, dissemination or distribution of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you. Tellabs ============================================================ _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers ============================================================ The information contained in this message may be privileged and confidential and protected from disclosure. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any reproduction, dissemination or distribution of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you. Tellabs ============================================================ _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From WHITERL at nv.doe.gov Wed Oct 29 11:23:06 2008 From: WHITERL at nv.doe.gov (White, Bob) Date: Wed, 29 Oct 2008 09:23:06 -0700 Subject: Summary: Solaris 10 Newbie - Revised Message-ID: Original problem: SMC could not find an SMC server. Attempts to stop and restart the server, and reset the server failed to clear the problem. I did call Sun Solve, and started a case #. Then played telephone tag for a bit. Continued to sift through all the threads that Google found on this topic while on hold. I found several mentions that this problem would be fixed in the next release. As the threads I was reading were from last fall and spring, and as the release that came out of the box was 8/07, I decided to just download the 5/08 release, and do a clean install. Finished that this morning. (Side note: I tried doing an upgrade, but the window environment wouldn't come up after reboot. Only the command line. I ran through the process again, this time doing a clean install and that worked better.) Got the new release running, and tried SMC again, and it works as advertised now. However - the consensus of all the responses I got was to eschew the SMC GUI and stick with the command line apps to do the system management tasks. A couple of responders (Dean Ross-Smith and Paul Ketelaar) suggested that I should look at Webmin, which I shall try to do. Thanks to all who responded so kindly. In the future I'll try to select a subject line that better reflects the problem. BTW: The only book reference I received was this one. Solaris 10 - The Complete Reference ISBN 0-07-222998-5 by Dr Paul Watters However, the reviews on Amazon.com were not very kind. Bob White M/S NLV075 Office: 702-295-2939 Fax: 702-295-2934 Cell: 702-630-0352 whiterl at nv.doe.gov [demime 1.01b removed an attachment of type image/gif which had a name of tech.gif] _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers From ghenry at cmi.univ-mrs.fr Fri Oct 31 09:18:28 2008 From: ghenry at cmi.univ-mrs.fr (Gerard Henry) Date: Fri, 31 Oct 2008 15:18:28 +0100 Subject: SUMMARY: smb on solaris 10 and macosx clients In-Reply-To: <48FF200A.9020604@cmi.univ-mrs.fr> References: <48FF200A.9020604@cmi.univ-mrs.fr> Message-ID: <490B13B4.7060203@cmi.univ-mrs.fr> thanks for all replies, but i have no solution for my problem. Changing the behaviour on the user's side isn't possible. Netatalk is great, but i don't dare to put it in a prod server. Here are the detailed proposed solutions: DRoss-Smith at reviewjournal.com wrote: > you could build netatalk > http://netatalk.sourceforge.net/ > > or buy ethershare > http://www.helios.de/products/ES/EtherShare.phtml Matthew Taylor wrote: > What do you mean by "interoperate"? I use Mac OS X (10.5) as my workstation > and manage several Solaris 10 boxes. I have not enabled SAMBA as I see no > need. My preferred solution for file exchange is to use Macfusion and > Macfuse (which see below), if I am not just doing one off sftp tranfers. > > http://www.macfusionapp.org/ > http://code.google.com/p/macfuse/ Rajiv Gunja wrote: > Look into FUSE and SSHFS it is secure and is per user than per server. Karl Vogel wrote: > "/" is never a legitimate character in a filename on a Unix box; it's > used strictly as a directory separator. You'll need to rename the > files on the Mac side if they contain "/". > i've enabled samba on my solaris 10 machine according: > http://blogs.sun.com/timthomas/entry/samba_and_swat_in_solaris > http://blogs.sun.com/timthomas/entry/enabling_and_configuring_samba_as > > on a macosx client, i can mount the share, and copy files into without > troubles, except when filenames contains characters like / for example > Enabling NFS isn't a solution for me, anybody knows if it is possible to > interoperate between macosx and solaris? _______________________________________________ sunmanagers mailing list sunmanagers at sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers