PDA

View Full Version : OnStream SCSI tape backup problems



Deckard
02-08-2001, 12:29 AM
I'm currently having an issue with one of my clients running ASIP 6.3.2 on a Yikes 400Mhz G4 and an OnStream Echo 30e external ADR tape backup drive connected to an Initio Miles Bluenote. Approximately every 3-5 nightly backups (at least once weekly), the backup fails with Retrospect's infamous 206 error (dirty heads, bad media, etc.). After buying an ADR cleaning tape, contacting both OnStream and Dantz, and switching out the drive for a new unit, the problem remains. Since OnStream's tech support is largely clueless about Mac issues and Dantz wasn't much help either, I've turned to the forum here to help as I bought the Bluenote card from MacGurus.

I've come to understand Retrospect's 206 error is really a generic communication problem and does not always mean dirty tape heads. After some experimentation, it bears no relation to use of the cleaning tape. I can have a bad backup one night, not clean the heads, and get a perfect backup the next. I've found no correlation with particular tapes causing the error. That leads me to believe it could be a SCSI issue with the Bluenote, the cable or otherwise.

A little background: I purchased the Yikes G4 in November of 1999 and the Bluenote shortly thereafter to provide legacy, bootable SCSI support to the G4. It was not known at that time a larger capacity backup device like the Echo 30e would be needed, nor did I know it would exist or be purchased. Backups of the server had been done with a 4GB DAT drive on a client computer with server volumes mounted. After realizing the DAT drive no longer had sufficient capacity, I researched other solutions and came upon the new OnStream drive. Although the drive was new, initial reviews were good and the price was hard to beat for the capacity. I was unaware of the upcoming Ecrix VXA drives in April of 2000 when the Echo 30e was purchased. Given my experience, I would gladly have purchased the Ecrix in its place. That's really not an option now.

Any ideas about this? I've scanned the SCSI Page of Doom here at MacGurus and thought perhaps some Bluenote settings might be the issue, but I'd like further advice. Sorry for the long post, just trying to be thorough.

System is as follows:

мя 400Mhz Yikes G4
мя 192MB RAM
мя AppleShare IP 6.3.2/Mac OS 9.0.4 with all current updates
мя Internal ATAPI Zip 100 drive
мя Initio Miles Bluenote SCSI card (latest 2.11 SIM, SmartSCSI 1.05)
мя OnStream Echo 30e SCSI tape drive (latest 1.08 firmware, only device attached, terminated, ID 4)
мя MiniDB50-MiniDB50 SCSI cable used, no adapters (included with drive)
мя Retrospect Desktop 4.3 (no clients, only server backed up)

Aproximately 8GB of data is backed up and compared nightly, about an hour for each. All programs except ASIP components are quit prior to backup (currently only FMP Server 3.0v4).

[This message has been edited by Deckard (edited 08 February 2001).]

Louie
02-08-2001, 12:58 AM
BlueNote settings jumped to mind as soon as I started reading your post. I've used lots of BlueNotes, but always with internal connections. I'm sure yours is external. Initio warns that for external connections, the synch rate should set down to 10 MB/s. We've argued amongst ourselves that you could get away with 20 MB/s with a good adapter, but we haven't tested it. Never-the-less, check your synch rate usind BlueNote 1.05. I would not use the cheap adapter that comes with card (I see you don't).

The card does Parity check as default, so set the recorder to match. Also, set the recoder to receive termination power from the SCSI bus. Of course, be sure you have good termination at the recorder and that cable is not too long and is of high quality (often furnished cables aren't of high quality). It might be wise to try other ID's for the recorder. If it's as dirty as a scanner, it might like 5 or 6 better.

All this is based on the assumption that the recorder is a SCSI-2 device. The Ecrix is LVD but will run Single-Ended.

Lots of luck. Some of the others may think of something else.

[This message has been edited by Louie (edited 08 February 2001).]

magician
02-08-2001, 01:03 AM
well...looking at your config, it seems about the only things left to try are a good Granite terminator on the rear of the Echo....(it does have two 50-pin MicroD connectors, right?)

part number GD8233 (http://www.macgurus.com/shoppingcart/obj_show_page.cgi?mgscsiterminators.html) should do it for you. You may also want to try part number GD3076 (http://www.macgurus.com/shoppingcart/obj_show_page.cgi?mgscsicables50microD.html) or some other Granite 50-pin MicroD cable, remembering that the shorter, the better.

once you've done that, all that remains is to talk to Initio to see if they have any guidance (doubtful, though worth an email).

is there a diagnostic firmware utility of some kind for the Echo that you can use to flash and test it? Any updates for it?

also...is the Echo supported directly by Retrospect, or did it require something from the ADK or a custom driver?

Deckard
02-08-2001, 02:56 AM
Thanks Louie and magician for the quick responses, I will try those suggestions.

Only thing connected to the Bluenote is the Echo 30, externally. Internal connector is not used.

The Echo 30 is using the latest 1.08 firmware, I've added that to my initial post.

The Echo 30 SCSI is supported directly by Retrospect 4.3 and the lastest Echo firmware. I believe only FireWire version of the drive is currently waiting on full support.

According to OnStream's specs the drive is SCSI-2, and the cable certainly is.

To be honest, I can't recall if the Echo has one or two MicroD50 connectors on back, I'll check next time I'm there. The Granite terminator and cable seem like good ideas. If setting the sync rate on the Bluenote and term power from the SCSI bus don't fix it, I'll try the Granite stuff next.

trag
02-08-2001, 02:17 PM
This is a long shot, but you might also want to check the time listed in the log files and see if the failures are happening at the same time when they occur.

I used to manage a back up system for a group of about 50 (PCs) in a 3000 person installation. We had our own Netware server on the larger LAN with a DAT drive attached to it. Backup was failing every night after backing up about half the users. I pulled a lot of hair, switched SCSI cards (a pain in a Netware/PC environment) and phoned tech support many times. Finally I noticed that the log indicated that the failure always occurred at 11:35 PM.

I checked with the network administrators but was never able to identify the root cause. In my case I ended reprogramming the backups so that half would finish before 11:30 and the other half would begin after 12 AM.

Anyway, you might just check the time in the log and see. That's probably not it, but it's one of those weird possibilities that it could take you a long time to think of.

magician
02-08-2001, 02:52 PM
now that HAS to be related to the tides and the phases of the moon!

http://macgurus.com/infopop/emoticons/icon_biggrin.gif

Deckard
02-15-2001, 08:22 PM
I've tried some of the suggestions here, but the problem still seems to be occurring. I checked with SmartSCSI and the Bluenote was already running at 10MB/sec. Parity is set to on with the Bluenote, but I see nowhere on the Echo to set this feature, either in software or on the drive itself. I changed Bluenote termination from single 50-pin to automatic, the only change I hadn't already made for the card.

I think it at least one case, the tape itself is bad. I tried manual backups three consecutive times with one tape and each time it failed. After checking the Retrospect log, I found that same tape had failed previously, even right after a cleaning.

I guess next things to try are the hidden SCSI prefs settings in Retrospect (SCSI Select with Attention, Blind SCSI Transfers, etc.) and going with Granite Digital cables. I did notice the included cable is at least 3 feet long, maybe 4, and it only needs to be 18" or so. One question about that: If I get the cable and the terminator, do I then turn off the termination switch on the tape drive? I'm pretty sure that's the way it works.

I'll let you know how it goes. Thanks again for the help.

Louie
02-15-2001, 09:20 PM
Don't double terminate.

magician
02-16-2001, 03:14 PM
right. In other words, yes. If you install an external terminator, you would turn the termination switch off.

http://macgurus.com/infopop/emoticons/icon_smile.gif

Deckard
03-23-2001, 10:52 AM
OK, finally got approval for ordering the GD SCSI cable and terminator you recommended magician, and did so last week. I'll let you know how it goes once they're installed.

As I'm sure you're aware by now, OnStream just filed for Chapter 7 bankruptcy, so future support for this piece of crap tape drive is "uncertain" at this time. Of course, their European center claimed they're trying to get investors together to continue support. Yeah, I'm sure that'll happen.

http://www.macgurus.com/ubb/dance.gif <--- Corporate Executive Bullshit Dance.

Given the history of such promises from other former Mac-supporting companies (Micro Conversions, 3dfx, SyQuest, etc., etc.), that seems "unlikely" at this time.

Sorry if I sound bitter, four hours sleep and the fact that I am bitter are causing that. http://macgurus.com/infopop/emoticons/icon_smile.gif

[This message has been edited by Deckard (edited 23 March 2001).]

ChrisYip
03-23-2001, 11:26 AM
That's a pain about Onstream...I've been a big convert to the VXA system - we have 3 tape drives in house now and they are working out really well - the bundled kits are nice and Ecrix's Customer Service is stellar.

Chris.

Deckard
05-04-2001, 11:13 AM
Well, maybe some miracles do occur. OnStream has risen from the ashes to once again be a player in the tape backup market.

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>This from today's Macintouch (http://www.macintouch.com):

When tape storage vendor OnStream Inc. closed in March, we reported that a group of former executives were working on a plan to resurrect the company with new financing and product plans. Edward Snelleman reports that a new company, OnStream Data, was founded on May 1, after acquiring "the intellectual property from OnStream Inc."<HR></BLOCKQUOTE>

On Wednesday I ordered the VXA FireWire drive from Ecrix on their $200 off special with 30-day MBG. I'm sure the SCSI version would be faster, but I'm not trusting SCSI right now on that server after the Onstream problem. With $200 off, the FireWire drive is only $50 more than the SCSI anyway (I don't need the Retropspect bundle) .

Unfortunately the Granite MicroD50 cable and terminator made no difference whatsoever. I'm eating their cost in order to keep my client happy.

At least OnStream's resurrection may get me more money for my client when selling the Echo30 ADR drive on eBay, and all those tapes.

[This message has been edited by Deckard (edited 04 May 2001).]

ChrisYip
05-04-2001, 01:05 PM
Hi there..let me know how that FireWire one works - it's supposed to be cross-platform too at least with WinMe and Win2K... My three LVD VXA's are running off a Miles2 and, well, have been relatively happy (see my discussion earlier on funky voodoo stuff related to OS 9.1 ... and booting ...)

Chris

Deckard
05-09-2001, 08:36 AM
Hi Chris,

So far the FireWire VXA-1 has been glorious simplicity itself, and hasn't missed a beat on the backups. I'm averaging 115 MB/minute overall (backup and verify), with no problems yet, and it's faster than the OnStream's Echo 30e SCSI-2 drive to boot. This thing is great! I wish Ecrix had come out with this six months earlier with this drive and I'd read more (or any) bad reports about the OnStream unit before purchase. No wonder Ecrix beat the hell out of Onstream in the market, even with the disparity in pricing. Quality is quality, and worth paying for (especially when you have clients willing to do so).

One question Chris, how often have you had to use the cleaning tape in the VXA-1? The manual says every 70 hours, but as manuals go, they can be far different than real world results. Just wondering how high the set the cleaning warning time in Retrospect (in hours) before the drive will tell you it's time to be cleaned. My nightly backup and verify runs just under two hours.

Thanks!

ChrisYip
05-09-2001, 10:59 AM
Excellent report - I'll put an order in for the FireWire one then.. thanks

As for the cleaning tape -well - I'm actually using Retrospect's default message of 20 hrs and running it that way. Our backups can take ~ 14 hrs onto two tapes over our 10 /100 network - wayyy too much backing up of lab stuff.

I'm probably overdoing it at that rate...

I'm back to having funky Retrospect network backing up problems again - odd Type 10 errors keep cropping in.. it's very odd.. I'm thinking that it might be my network switches that are discombobulating everything

Chris.

Deckard
05-12-2001, 08:53 PM
Still rocking after a week of testing on the FireWire VXA-1. Not even a burp. We're going with the full-on purchase of the tapes we need now.

I'll let you know what I find on the MTBC (Mean Time Between Cleanings) for the li'l hummer. Hasn't been enough time yet to need one.

Now I'm just waiting for OnStream to crank back up, supposedly this month, so I can see if they'll honor the "lifetime" warranty on the bad tapes. If they do, I'll be flicking that Echo30e booger off on ebay as soon as I can. I wouldn't sell it without proper warning on the ebay listing about OnStream's troubles, of course. I'm kinda honest like that.

magician
05-14-2001, 04:53 AM
no need to eat the cost. If you're still within 30-days, just talk to Deb. She'll take care of you.

http://macgurus.com/infopop/emoticons/icon_smile.gif

Deckard
05-31-2001, 02:49 PM
Not a big deal on the cable, I'll keep it for my busines and compensate my client accordingly.

The FireWire tape drive is still humming along after many backups to the 12MB sized tapes, not a single problem yet. We're at about 25 hours use and still no need to use the cleaning tape yet. The Ecrix VXA-1 is awesome!

I'm dancing again! http://www.macgurus.com/ubb/dance.gif

Deckard
07-08-2001, 09:59 AM
Hi all. I finally got the first error in Retrospect that the new Ecrix FireWire needed to be cleaned. It was at about 63 hours, the first time I've had to run it. I'll post again later so we can perhaps get an average. The backup now runs nightly to a second internal hard drive, so the tape backup is simply for offsite storage.

This is opposed to the 3-5 times weekly that Retropsect reported the Onstream Echo 30 needed it. Last response I got from Onstream was that I would need to return the bad tapes to my original vendor, then they could get new tapes from Onstream. In other words, no direct replacement for bad tapes from the Onstream "shell" company they reassembled here in the U.S. I hope that changes sometime, I'll be the original web vendor would laugh in my face about taking them back.

I'll post if I get any good results on returning the bad tapes.

ChrisYip
07-08-2001, 10:23 AM
is that the setting that Retrospect has in its prefs?. I think mine is set at 24 hrs or something like that....

I routinely throw the VXA cleaning tape into our machines once a month just to be safe - it's probably over kill though.. our main backup runs about an hour a day, with the network one taking about 4 hrs...once a week..

Chris

Deckard
07-09-2001, 09:03 AM
Hi Chris. Indeed it is that setting I used. I checked the log and found that Retrospect gave a 206 error at around 63 hours. I had the setting higher than that, so I'm sure it's accurate. Further testing will bear that out as and average MTBC (Mean Time Between Cleanings), I hope.

The drive is located in a closet with relatively low dust (a wall fan sucks heat and dust out), so your mileage may vary if you have more dust in your environment. I suggest testing this yourself on one system just for grins. As long as you have good backups, missing a single one when Retrospect gives the 206 error shouldn't be a problem. Hope you find this useful. http://macgurus.com/infopop/emoticons/icon_smile.gif