asv@gaboon.UUCP (Stan Voket) (07/06/90)
The more I read this group the more nervous I get!
I've got my 2.2 upgrade still sitting here on the floor in the shipping
box. I've got one machine here and can't spend 3 weeks of downtime
doing this upgrade.
Is there anyone out there who has upgraded form 2.0.2 to 2.2 and
encountered no problems? I'm ready for some moral boosting good news!
I think a a few positive "upgrade testimonials" would be generally
useful and welcome here. (I hope they exist!)
Has anyone done the upgrade rather than the destructive reinstall on a
big ESDI disk? Any problems with mail and news not working?
Thanks! :-)
--
+----------------------------------------------------------------------+
| - Stan Voket, asv@gaboon - OR - ...uunet!hsi!stpstn!gaboon!asv - |
| - Voice: (203) 746-4489 - FAX 746-9761 - |
+----------------------------------------------------------------------+cpcahil@virtech.uucp (Conor P. Cahill) (07/07/90)
In article <6417@gaboon.UUCP> asv@gaboon.UUCP (Stan Voket) writes: > Has anyone done the upgrade rather than the destructive reinstall on a > big ESDI disk? Any problems with mail and news not working? I did the upgrade on a 680MB ESDI drive and had just a couple of problems ( like some of my crontab entries got blown away, my gettydefs file got blown away,etc). If you have room, make sure you have an on-line backup copy of your root FS so that you have immediate access to whatever gets blown away. BTW - by blown away I mean that the file on the update disk replaces the equivalent pre-existing file. We installed 2.2 and had no problems with most of the stuff that had already been configured on the system (mail, news, uucp, even the pre-existing kernel config stuff). Our only complaint (other than the #! bug that I reported before) is that our Archive 60MB tape no longer works (using either the new driver included with 2.2 or with the old driver that worked under 2.0.2). ISC is looking into the problem. Otherwise the system is running great. We have seen no chane in percieved performance (i.e. it hasn't changed, but we haven't run a benchmark to verify it). All of our other peripherals (including WD ethernet, Sigma Designs Laser View Display, DPT caching controller, and Equinox Megaport) work flawlessly. 2.0.2 and 2.2 co-exist on the same network without any problems. -- Conor P. Cahill (703)430-9247 Virtual Technologies, Inc., uunet!virtech!cpcahil 46030 Manekin Plaza, Suite 160 Sterling, VA 22170
jackv@turnkey.tcc.com (Jack F. Vogel) (07/07/90)
In article <1990Jul07.125218.24439@virtech.uucp> cpcahil@virtech.UUCP (Conor P. Cahill) writes: >In article <6417@gaboon.UUCP> asv@gaboon.UUCP (Stan Voket) writes: >> Has anyone done the upgrade rather than the destructive reinstall on a >> big ESDI disk? Any problems with mail and news not working? > >I did the upgrade on a 680MB ESDI drive and had just a couple of problems ( >like some of my crontab entries got blown away, my gettydefs file got blown >away,etc). If you have room, make sure you have an on-line backup copy >of your root FS so that you have immediate access to whatever gets blown >away. I have a WD1006 based system rather than ESDI, and also did the upgrade rather than reinstall. I also had no other problems than what Conor mentioned. Just remember to make backup copies of your crontab files, /etc/inittab, and so forth. I don't know whether you want to keep the old gettydef since it looked like they had changed it for enhanced functionality but you will probably want to keep an old copy for reference. Oh, I also did not install the BNU upgrade, but if you do I would keep a backup of Permissions, Systems, Devices, Dialers. >We installed 2.2 and had no problems with most of the stuff that had already >been configured on the system (mail, news, uucp, even the pre-existing kernel >config stuff). ditto, had the upgrade done in a couple of hours, brought the system back up, and uucp and news restarted as though nothing had happened. The only thing that ISC should fix is the crontab overwrite. Great job guys!! >Our only complaint (other than the #! bug that I reported before) is that our >Archive 60MB tape no longer works (using either the new driver included >with 2.2 or with the old driver that worked under 2.0.2). ISC is looking >into the problem. This is the main reason I did the followup on this. This is real strange Conor, when I saw this I realized that I had not even tried the tape since the upgrade and my heart skipped a beat :-} (I also have an Archive 60M, not the SCSI). So I shoved in one of my cpio backup tapes, did a listing and it worked just fine, hmmmmm?!? How does yours not work? I seem to remember that the /etc/conf/sdevice.d/ct file had been overwriten and the interrupt was set back to 3 so I edited it and changed it back to 5 where mine is set, did you check this?? Let me know how things turn out. Disclaimer: As always, these are my views, not my employer. -- Jack F. Vogel jackv@locus.com AIX370 Technical Support - or - Locus Computing Corp. jackv@turnkey.TCC.COM
larry@nstar.uucp (Larry Snyder) (07/07/90)
asv@gaboon.UUCP (Stan Voket) writes: > Is there anyone out there who has upgraded form 2.0.2 to 2.2 and > encountered no problems? I'm ready for some moral boosting good news! Yes.. No problems at all - and I did a complete install from scratch then restored the tapes.. -- Larry Snyder, Northern Star Communications, Notre Dame, IN USA uucp: iuvax!ndmath!nstar!larry -or- larry@nstar Public Access Unix Site (219) 289-3745 / lots of files & free PEP feeds!
cpcahil@virtech.uucp (Conor P. Cahill) (07/08/90)
In article <1990Jul07.152106.23429@turnkey.tcc.com> jackv@turnkey.TCC.COM (Jack F. Vogel) writes: > >>In some article I (Conor) wrote: >>Our only complaint (other than the #! bug that I reported before) is that our >>Archive 60MB tape no longer works (using either the new driver included >>with 2.2 or with the old driver that worked under 2.0.2). ISC is looking >>into the problem. > >This is the main reason I did the followup on this. This is real strange >Conor, when I saw this I realized that I had not even tried the tape since >the upgrade and my heart skipped a beat :-} (I also have an Archive 60M, not >the SCSI). So I shoved in one of my cpio backup tapes, did a listing and it >worked just fine, hmmmmm?!? How does yours not work? I seem to remember that >the /etc/conf/sdevice.d/ct file had been overwriten and the interrupt was >set back to 3 so I edited it and changed it back to 5 where mine is set, >did you check this?? Let me know how things turn out. Before I go in to our symptoms I just want to say that Yes I checked and verified that the tape drive is configured exactly as it was before (I made a backup copy of / in a another file system which was available for comparison). Anyway, the symptoms are as follows: We can place a tape into the drive and run any of the ctape commands like retension or rewind without any problems (the tape does go through the indicated operation). When we attempt to write data to the drive, it goes through the first few moves of the tape, stops (tape light is still on), and the process is locked waiting for the operation to finish. The same thing happens on a read of a previously made tape. We have tried different tapes both reading and writing, different drivers (I still had the old driver lying around). The hardware configuration did not change a bit from 2.0.2. It is possible that the tape drive decided to go bad at the moment that the system was upgraded, but I really doubt it. -- Conor P. Cahill (703)430-9247 Virtual Technologies, Inc., uunet!virtech!cpcahil 46030 Manekin Plaza, Suite 160 Sterling, VA 22170
tim@comcon.UUCP (Tim Brown) (07/09/90)
In article <6417@gaboon.UUCP>, asv@gaboon.UUCP (Stan Voket) writes: > The more I read this group the more nervous I get! > Is there anyone out there who has upgraded form 2.0.2 to 2.2 and > encountered no problems? I'm ready for some moral boosting good news! Ok, credit where it is due! I have installed 2.2 on three systems. I will list them and their problems that remain. 1) 386/20 with 8 megs RAM AMI BIOS C&T Set MB 150 meg SCSI/Future Domain 865 Controller with floppy support. This was an upgrade type install. I did it this way because I didn't have much time to do the upgrade and it seemed to be the safest approach. It worked well for the most part. Probelms: sendmail will not die if TCP is not properly installed. Even if I switch to init level 2 and get rid of the hosts file *and* I had sendmail configured to *only* use uucp for mail the whole time. If I set up a complete TCP network sendmail functions fine. This system is *definitly* slower than it was before the upgrade. Even after the switch to init level 2, it is slower. I am fielding complaints of slowness from users all day. I am seriously thinking of switching this one back to 2.0.2. The main reason for the upgrade on this machine was for TCP, when that is running the machine craaawlss. NOTE: Other than the above mentioned problems the system works fine. Not having a support contract however (and not going to buy one!) nor hearing any answers here, I will probably *have* to switch back to 2.0.2. 386/20 8 megs RAM AMI C&T MB Adaptec 2372B Controller with 2 Seagate 4144R's This was a destructive re-install. This system runs *fast* and perfectly, with the exception of the sendmail bug mentioned above. This system has the news on it. I would *like* to run SLIP here if I could figure out how to do it. 386/33 12 megs RAM Micronics MB, Pheonix BIOS WD1007wa2-wah controller, and 2 micropolis ESDI drives (don't remember the exact model) 330 megs each. This one was a destructive reinstall. This system runs flawlessly. This is my favorite machine. Seems faster than before. I have TCP running, PCI and sendmail works right. There is one minor problem they all now experience with paging which I will post seperatly. Tim Brown nstar!comcon!tim
cpcahil@virtech.uucp (Conor P. Cahill) (07/09/90)
In article <1990Jul07.231614.26267@virtech.uucp> cpcahil@virtech.UUCP (Conor P. Cahill) writes: > > [my sad story about dead tape drive following upgrade to 2.2 deleted] > Aparently what has happened is that my tape drive decided to kick the big one just as I upgraded to 2.2. This would be a bit too coincidential for me if I hadn't (just now) taken the tape drive out of my machine and loaded it onto a 2.0.2 machine and it still had the same problem. Oh well, back to the vendor to get a replacement (I bought it just under 1 year ago). -- Conor P. Cahill (703)430-9247 Virtual Technologies, Inc., uunet!virtech!cpcahil 46030 Manekin Plaza, Suite 160 Sterling, VA 22170
gary@sci34hub.UUCP (Gary Heston) (07/09/90)
In article <1990Jul07.152106.23429@turnkey.tcc.com> jackv@turnkey.TCC.COM (Jack F. Vogel) writes: >In article <1990Jul07.125218.24439@virtech.uucp> cpcahil@virtech.UUCP (Conor P. Cahill) writes: >>like some of my crontab entries got blown away, my gettydefs file got blown >>away,etc). If you have room, make sure you have an on-line backup copy >>of your root FS so that you have immediate access to whatever gets blown >>away. >I have a WD1006 based system rather than ESDI, and also did the upgrade >rather than reinstall. I also had no other problems than what Conor mentioned. >Just remember to make backup copies of your crontab files, /etc/inittab, and >so forth. .... In order to save space, and ease disaster recovery, you can create a skeleton image of the root fs directories, and link the configured files (inittab, crontabs, uucp device files, passwd, and so on) into it. Then, make a quick backup of this directory, do your update (which will probably overwrite a few files, no matter which vendor you're dealing with), and restore the directory. Since the files will still be linked, restoring the directory restores the configuration, and saves lots of hair pulling. For example, I use a directory named /x for this, so I have /x/etc, /x/bin, /x/usr/bin, /x/usr/lib, and whatever subsidiary directories I need. Then: ln /etc/passwd /x/etc/passwd ln /etc/inittab /x/etc/inittab ln /etc/gettydefs /x/etc/gettydefs ..... (whatever other files you configure...) (This can be a cron-executed script to make sure things don't get accidentally unlinked.) Now, if you tar -cvf /dev/rfd0 /x everything gets saved. The tar can be done from within /x if you want to restore directly to the system files, by doing the tar -x from / instead of /x. The only hangup is that this only works within a file structure, obviously, so if you have /usr mountable you have to save those files separately. I came up with this trick while testing successive beta releases, to save myself a lot of re-config work. It's proven to work quite well, and eliminates the need for a separate root fs image. It saves backup time since you don't need to dump the whole root fs, just the one directory. Comments? -- Gary Heston { uunet!sci34hub!gary } System Mismanager SCI Technology, Inc. OEM Products Department (i.e., computers) "The esteemed gentleman says I called him a liar. That's true, and I regret it." Retief, a character created by Keith Laumer.
howardl@wb3ffv.ampr.org ( WB3FFV) (07/10/90)
From article <6417@gaboon.UUCP>, by asv@gaboon.UUCP (Stan Voket): > > The more I read this group the more nervous I get! > encountered no problems? I'm ready for some moral boosting good news! > > I think a a few positive "upgrade testimonials" would be generally > useful and welcome here. (I hope they exist!) Hello Stan, Well I am not sure if this is what you want to see, as I started out with the 2.2 Workstation Developer. I almost started with 2.0.2, but new I would be spending additional $$$ for an upgrade in just a matter of weeks. I did actually convert from AT&T V/386 Release 3.2 over to Interactive since they supported TCP/IP and NFS which I will need soon, plus they were suppose to have a much faster file system (I have seen some improvement). Anyway I have an 80386 system which is using the DTK PEM-3300 (33mhz) motherboard with the DTK bios installed. The system has 64K cache and 8 MB os main memory (thinking of going to 12). The drive is a Miniscribe 9780E (780meg) ESDI disk with is being used with the DPT Smart-Cache disk controller. I am also using the Comtrol Smart-Hostess 8 port serial controller, and the Adaptec VP150i (150meg) tape backup unit. The viseo display is the ATI VGA-Wonder 512 and it is being used with the Princeton ULTRA-16 monitor. I had no problem with the 2.2 instalation, and have everything installed except the 2K file system which I did not need. I can't say how well the TCP/IP and NFS system will work, as I have had the time to get it running yet (but hope to soon). I was very concerned about the 2.2 system, since I saw the bad postings regarding the DTK machines (but they were on the 25mhz boxes), but since I had already ordered the 2.2 system from Interactive I figured I might as well give it a shot. Anyway, to make a long story short, I have managed to get the Interactive UNIX Release 2.2 Workstation Developer up and running without a single problem. As of this posting it has been running for about two weeks, and it looks like it probably stable (I hope :-). ------------------------------------------------------------------------------- Internet : howardl@wb3ffv.ampr.org | Howard D. Leadmon UUCP : wb3ffv!howardl | Advanced Business Solutions TELEX : 152252474 | 210 E. Lombard St - Suite 410 Telephone : (301)-576-8635 | Baltimore, MD 21202
cpcahil@virtech.uucp (Conor P. Cahill) (07/10/90)
In article <690@sci34hub.UUCP> gary@sci34hub.sci.com (Gary Heston) writes: > > [description of using linked copies of config files to save backup time] > [and ease re-installation following an update deleted ] > >I came up with this trick while testing successive beta releases, >to save myself a lot of re-config work. It's proven to work quite >well, and eliminates the need for a separate root fs image. It saves >backup time since you don't need to dump the whole root fs, just the >one directory. An easier mechanism is to do the following: Once you have completed the original installation of your software create a file called ".epic_file" in your root directory. When you want to make a back up of the changed files you just run: find . -newer .epic_file -print | cpio ... This way you don't have to go through the step of creating the links and you don't have to worry about an update installation that removes the pre-existing files instead of copying over them. If you didn't create the ".epic_file" when you originally installed the software, you can use "touch MMDDHHMMYY .epic_file" to creat it with the appropriate date/time. -- Conor P. Cahill (703)430-9247 Virtual Technologies, Inc., uunet!virtech!cpcahil 46030 Manekin Plaza, Suite 160 Sterling, VA 22170
cpcahil@virtech.uucp (Conor P. Cahill) (07/12/90)
In some article, I sat here telling a sad story about a dead Archive tape drive > Before I go in to our symptoms I just want to say that Yes I checked and > verified that the tape drive is configured exactly as it was before (I made > a backup copy of / in a another file system which was available for > comparison). Ah, I check the sdevice file, but not the mdevice file. Boy do I feel like a fool (but I'm also mad at ISC - they shouldn't have overwritten my mdevice entry). Anyway, the problem is that the ISC update changed the DMA channel from a 3 to a 1 and hence the problem. Thanks go out to Jim Sullivan of HCR (now SCO, Canada), who describe the problem exactly. -- Conor P. Cahill (703)430-9247 Virtual Technologies, Inc., uunet!virtech!cpcahil 46030 Manekin Plaza, Suite 160 Sterling, VA 22170