INFO-VAX Tue, 01 Jul 2008 Volume 2008 : Issue 364 Contents: Re: Backup to SnapServer Re: DEGPA in DS20 shows "offline" Re: Expanding a RAID5 array Re: Expanding a RAID5 array Re: Expanding a RAID5 array Re: Expanding a RAID5 array Re: Expanding a RAID5 array Re: Expanding a RAID5 array Re: It would be nice to have on VMS OT: Disturbing thoughts on creation of the universe ---------------------------------------------------------------------- Date: Mon, 30 Jun 2008 21:10:05 -0500 From: David J Dachtera Subject: Re: Backup to SnapServer Message-ID: <486991FD.3C0ADB3C@spam.comcast.net> Carmine Castiglia wrote: > > Can anyone offer some tips on how to (if even possible) backup and restore > drives from an Alphaerver running OpenVMS v7.1-1H2 across an ethernet > connection to a SnapServer 4100 (already on the same network)? Well, not knowing what a "SnapServer" is, I'd have to offer the usual advice: be careful when using anything that is filesystem-agnostic. Bit-for-bit transfers can be useful, but can also be very tricky. D.J.D. ------------------------------ Date: Tue, 01 Jul 2008 02:55:27 GMT From: Malcolm Dunnett Subject: Re: DEGPA in DS20 shows "offline" Message-ID: Len Whitwer wrote: > > It's remotely possiable that your card has gone bad. Yes, it turned out to be a bad card (Perhaps it got damaged in the move from the DS10L to the DS20E). I swapped in another card and it's fine now. ------------------------------ Date: Mon, 30 Jun 2008 19:59:38 +0100 From: baldrick Subject: Re: Expanding a RAID5 array Message-ID: Len Whitwer wrote: >> On Jun 28, 8:19 am, "Carmine Castiglia" wrote: >> >>> I am running an AlphaServer 1200 with a RAID Array 230/Plus subsystem >>> installed. This question concerns a RAID 5 array consisting of three >>> DS-RZ1CB-VW (4.3GB) drives. > data sets using the 18.2GB disks is much faster than a RAID5 > but not as efficeint > with space) "18GB drives are very cheap these days" 18 GB not compatible with SWXCR......... -- nclews at csc dot com aka Mr. CP Charges ------------------------------ Date: Mon, 30 Jun 2008 15:59:21 -0400 From: "Carl Friedberg" Subject: Re: Expanding a RAID5 array Message-ID: <890539d90806301259x34d31b82l1bff2ad0ca0d5b4f@mail.gmail.com> In my experience, the SWXCR (2.13 firmware? it's been 10 years, memory is unreliable) had a 32 Gbyte limit, but it was happy to build 18GB raid 1 (mirror) from a pair of 18 Gbyte drives (blue bricks IIRC). Again a specific BA35x with appropriate narrow personality module is required to get this all working. I second, third or fourth whoever said to get an HSZ device. To do any kind of recover on the SWXCR, you have to shut down VMS and bring up the Alpha console (music, drumrolls, and VERY LONG WAIT) On Mon, Jun 30, 2008 at 2:59 PM, baldrick wrote: > Len Whitwer wrote: > >>> On Jun 28, 8:19 am, "Carmine Castiglia" wrote: >>> >>>> I am running an AlphaServer 1200 with a RAID Array 230/Plus subsystem >>>> installed. This question concerns a RAID 5 array consisting of three >>>> DS-RZ1CB-VW (4.3GB) drives. > >> data sets using the 18.2GB disks is much faster than a RAID5 >> but not as efficeint >> with space) "18GB drives are very cheap these days" > > 18 GB not compatible with SWXCR......... > > -- > nclews at csc dot com aka Mr. CP Charges > ------------------------------ Date: Mon, 30 Jun 2008 15:32:28 -0500 From: Chris Scheers Subject: Re: Expanding a RAID5 array Message-ID: baldrick wrote: > > Also if you have a spare configured, you need to make that the same as > the largest disk, it will also work for the smaller disks. And beware > narrow and wide disk differences. As far as I'm aware you can't get > narrow 9 gig disks, but you can get narrow 4.3 gig. You may need to > replace the shelf with a wide compatible BA356 and a narrow personality > module. FWIW: A RZ40 is a narrow 9GB drive. They do seem to be a bit rare. A BA356 with a narrow personality module and all the correct backplane jumpers will allow wide drives to be used with a narrow HBA. -- ----------------------------------------------------------------------- Chris Scheers, Applied Synergy, Inc. Voice: 817-237-3360 Internet: chris@applied-synergy.com Fax: 817-237-3074 ------------------------------ Date: Tue, 01 Jul 2008 00:00:16 +0100 From: baldrick Subject: Re: Expanding a RAID5 array Message-ID: Chris Scheers wrote: > baldrick wrote: >> >> Also if you have a spare configured, you need to make that the same as >> the largest disk, it will also work for the smaller disks. And beware >> narrow and wide disk differences. As far as I'm aware you can't get >> narrow 9 gig disks, but you can get narrow 4.3 gig. You may need to >> replace the shelf with a wide compatible BA356 and a narrow >> personality module. > > FWIW: > > A RZ40 is a narrow 9GB drive. They do seem to be a bit rare. > > A BA356 with a narrow personality module and all the correct backplane > jumpers will allow wide drives to be used with a narrow HBA. Yes but I don't believe the SWXCR supports any drive size above 4.3 GB. My 9.1 GBs were not even recognised as being present in the shelf. I have had 18.2 GB disks working in narrow mode (but not on a SWXCR!) I've worked test scenarios with these controllers, swapping 1 GB for 2.1 GB for 4.3 GB in both narrow and wide configurations. Mixing and matching doesn't seem to work either. Having said that, they are quite resistant to data loss (when you employ a protection system). Thanks for the mention of the RZ40, a Fujitsu re badged drive unless I'm mistaken. nic -- nclews at csc dot com aka Mr. CP Charges ------------------------------ Date: Tue, 01 Jul 2008 00:18:23 +0100 From: baldrick Subject: Re: Expanding a RAID5 array Message-ID: Carl Friedberg wrote: > In my experience, the SWXCR (2.13 firmware? it's been 10 > years, memory is unreliable) had a 32 Gbyte limit, but it > was happy to build 18GB raid 1 (mirror) from a pair of > 18 Gbyte drives (blue bricks IIRC). Again a specific > BA35x with appropriate narrow personality module > is required to get this all working. > My scenario involved multiple drives presented already. Did you configure in an 18 GB spare too? I must admit I didn't try too hard with very large drives (9 GB) it was overkill for the space requirements at the time, and the presence of other drives probably prevented it from working. I'd guess you have a 4-6 hour build time on that disk? The large presentations were raid5 only, the other disks were mirrors. I've used 9 GB on HSD10's in mirrors, never tested 18 GB, but I could... Its a bit of a waste because the SWXCR is a 3 channel controller to at least give you some performance (with its peers). -- nclews at csc dot com aka Mr. CP Charges ------------------------------ Date: Tue, 01 Jul 2008 07:45:01 +0300 From: =?ISO-8859-1?Q?Kari_Uusim=E4ki?= Subject: Re: Expanding a RAID5 array Message-ID: <4869b650$0$2662$9b536df3@news.fv.fi> Carmine Castiglia wrote: > I am running an AlphaServer 1200 with a RAID Array 230/Plus subsystem > installed. This question concerns a RAID 5 array consisting of three > DS-RZ1CB-VW (4.3GB) drives. > > I would like to swap this out with DS-RZ1DB-VW (9.1GB) drives and would like > to do so in the easiest, safest, and quickest way possible. I know that if > I replace a single 4.3GB drive with a 9.1GB drive and allow the array to > rebuild, the total storage space available will not change. This rule > applies if I replace a second drive as well because available storage space > is always calculated based on the smallest drive in the array. Correct? > > So, what happens if I continue the process and replace the third drive. > After rebuild has completed, will the RAID controller automatically make the > full (18.2GB RAID 5) capacity available? WIll this process be completed > without data loss? > > Here you can find the specifications for the KZPAC (RAID Array 230/Plus). http://h18002.www1.hp.com/products/quickspecs/SOC/QB00E7PF.PDF There you'll find the supported disks and the possible configurations. I tend to recommend building a Raid 0+1 set (with as many disks as possible) and when it is ready, to copying the data from old LUNs to newly built LUNs using BACKUP/IMAGE. That way you'll get most disk space with best performance and the downtime will take only as long as the Raid 0+1 build and the BACKUP will take. The 9GB Ultra disks are the fastest, but using more 4.3GB disks will give you a somewhat faster set, because there are more spindles. Using BACKUP for copying the data will also defragment your data as an additional benefit. Regards, Kari ------------------------------ Date: Mon, 30 Jun 2008 11:40:46 -0700 From: Marty Kuhrt Subject: Re: It would be nice to have on VMS Message-ID: Arne Vajhøj wrote: > Main, Kerry wrote: >>> From: Arne Vajhøj [mailto:arne@vajhoej.dk] >>> John Smith wrote: >>>> http://www.sybase.com/detail?id=1056945 >>>> First time I used Sybase IQ was about 10 years ago for a 2Tb initial >>>> load >>>> data warehouse I designed for a bank. A really great product and >>>> waaay fast. >>>> That one was on Solaris. >>>> >>>> I don't know of anything comparable to this product on VMS. > > >>> Sybase dropped VMS support a long time ago. >>> >>> The only major database vendor that supports VMS is Oracle. >>> >>> Those that want something like Sybase IQ should check out what >>> Oracle has. > >> Sybase has not done to much in the enterprise space in the last few >> years. They were focussing a lot on distributed DB market. Perhaps >> they are re-focussing on the enterprise again and are using this >> benchmark to jump start that marketing. >> >> In the mid to high enterprise DB space I usually see Oracle, SQL >> Server or DB2. > > Sybase is not as big a player as they used to be, but there are > still some very big Sybase sites out there. > > Arne Sybase used to be spread out through most of the buildings in Emeryville (CA, wedged between Berkeley and Oakland) at one point. They kept downsizing, and moving out of buildings. At one point it was a ghost town of empty buildings with remnants of Sybase signs on the side. They eventually moved out to a newer, smaller, closer to the C-level execs homes, office park in Pleasanton. It was handy for other startups in the area because they could get space with small data centers and high speed hookups. Sendmail, OFoto, MobiTV and Expressions (digital media training) all resided in ex-Sybase buildings at one point. I still see one of the Sybase guys that did VMS stuff way back when at computer geek parties on occasion. Every time I see him I ask, "Still working at Sybase?". "Yep", he'll say. "They dropped VMS support, they're DEAD to me!", I smirk. We'll both laugh a sad laugh and head back to the keg for a refill. The death of Sybase on VMS was pretty much the death of VMS for the main production stuff at Barra, the financial software company I worked for at the time. ------------------------------ Date: Tue, 01 Jul 2008 01:17:42 -0400 From: JF Mezei Subject: OT: Disturbing thoughts on creation of the universe Message-ID: <4869bdfc$0$30363$c3e8da3@news.astraweb.com> This is VERY OT. But so strange I can't resist posting it. Person X,male, aged 32, travels back in time 32 years. Person X has sex with his own mother, she gets pregnant and gives birth to ... Person X. Person X becomes his own father. And his son is himself. The baby would *have* to be an exact clone of his father since they are one and the same. When the egg is fertilised, the mixture of genes would occur in an absolutely predicted way where the female genes would exactly replace identical genes coming from the father (and those genes being replaced with the mother's genes originally came from the father's mother). There is an interesting portion here that the genetic mixup during fertilisation would be fully known in advance and predictable. This means that when the father has sex with the mother, the mother's genes would only replace the portion of the father,s genes which had come from his mother, producing an exact replica. Ok, sorry if this caused a few people up upchuck. But I am getting somewhere: Person X's genealogical tree would be interesting since all ancestors on the father's side would be born on the same date and in fact be the same person. The person would be his own father, and be his father's father etc etc. Infinite loop. At the time of insemination, while the result would be fully predictable (so much for eisenburg), the reaction would be physically normal. Sperm with certain human genetic makeup would mix with an ova with certain genetic makeup and create an offspring. However, in the big picture, the male portion of the genes would have never been created. It would merely exist after a certain point in time (the birth date of X). The genetic legacy might continue if X has more than 1 baby (perhaps before he travels back in time, he mates with a different female). But the genetic code would not exist prior to his birth. So in essence, a genetic identity would have been created out of thin air, or just merely exist. Perhaps a similar logic would explain the universe. Perhaps the "big bang" is just our universe travelling back in time a few billion years and when it rematerialises, it happens in a process similar to the big bang. If baby universe is created by a much older version of itself travelling back in time, it would mean that no "god" would be needed to create the universe in the first place since the universe would simply exist in a loop where it creates itself over and over again. And the matter/energy would essentilly be like a perpetual machine where all of it is recycled when moved back in time to the big bang. The matter/energy would merely exist in time, having never been created. ------------------------------ End of INFO-VAX 2008.364 ************************