PSA: Bad VirtualCD update, DO NOT INSTALL

H+H Software GmbH pushed out an update for VirtualCD 9 to v9.3.0.1 today.

DO NOT APPLY THIS UPDATE.

VirtualCD 9 will not run after updating to v9.3.0.1, and the bad installation cannot be fixed by doing a repair installation.  You will be able to fix your VirtualCD installation only by completely uninstalling the programs (you do not need to delete your configuration data) and performing a clean reinstallation of v9.3.0.0 or earlier.

New-to-us hardware, and svccfg questions

We’ve been gifted with some new hardware over the past year, including a rather nice 3U rackmount server that I’ve not been able to configure until the past few days (barring an abortive fight with FreeBSD¹) thanks largely to 3Ware², who — as previously mentioned — strung me along for nine months waiting for them to release Solaris 10 drivers from a 9500S SATA-RAID controller that they’d end-of-lifed about a month after I first inquired about the driver.

Anyway, last week, darthgeek (who gave us the machine in the first place) just sent me a pair of LSI 3080X SATA RAID controllers, and I’ve finally been able to install Solaris 10 and bring it up.  It’s a dual 3.2GHz Xeon box with 4GB of RAM (and will take up to 16GB in 2GB DIMMs), with two 80GB mirrored SATA boot disks, twelve 300GB SATA disks spread across those two LSI controllers, and three redundant power supplies.  This is replacing a Sun Ultra30 with a single 250MHz UltraSPARC II processor, 768MB of RAM, and a total of 414GB of disk, 360GB of it in two external 711 disk cases.  It’s also replacing one of my two Sun Ultra5s, and offloading our webserver, primary mailserver, MySQL server, and main backup server from my personal workstation, babylon5, as soon as I get everything configured.  Once I had the new disk controllers installed and the drives cabled to them, it took me about 24 hours to go from bare metal and never having run Solaris 10 before, to having babylon4³ up and running under Solaris 10 with the full storage array online (3.25TB of RAIDZ2), shared via NFS and Samba, with the entire contents of minbar‘s array copied over, and a new primary nameserver running as an unprivileged user in its own zone.

Some of that configuration is proving to be a little of a headache, though, because of issues with svccfg that I don’t fully understand yet, where the behavior of svccfg — a tool that configures system services — doesn’t match what’s documented in its man page, and does unexpected things.  So, if you’re a Solaris 10 export and know the inner workings of svccfg, please drop me a line...

[1]  Don’t ask.  I have never had so many kernel panics in such a short time in my life.

[2]  3Ware are asshats and wankers.  The least they could have done was actually tell me when they EOL’d it, rather than stringing me along until the driver finally shipped, nine months late, and only then and only when I specifically asked whether the 9500S was supported telling me that no, the new driver release didn’t support my 9500S and there never would be an updated driver for it and they’d EOL’d it eight months previously.

[3]  It’s our second machine to bear the name.  The first was a Sun Ultra Enterprise 3000 with 2GB of RAM, six 336MHz UltraSparc II processors and an internal rack full of 9GB Seagate Cheetahs.  It drew so much electrical power we couldn’t afford to run it except for testing and study.