Mailing List ArchiveSupport open source code!
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]Re: Linux 2.0.0
- To: tlug@example.com
- Subject: Re: Linux 2.0.0
- From: turnbull@example.com (Stephen J. Turnbull)
- Date: Mon, 10 Jun 96 12:15 JST
- Cc: tlug@example.com
- In-Reply-To: <199606100040.JAA14446@example.com> (message from Leo Glynn on Mon, 10 Jun 1996 00:40:35 +0000)
- Reply-To: tlug@example.com
- Sender: owner-tlug@example.com
>>>>> "Jim" == Jim Tittsler <jwt@example.com> writes: Jim> I've just installed Linux 2.0.0. It boots! (An improvement Jim> over the 1.2.0 release, if I recall correctly.) I've just Jim> used it to do another latex2html pass over my latest Jim> document. It works! Clearly it is stable and everyone Jim> should upgrade immediately. :-) I've been running pre2.0.2 for ten days now, no problems. I had no problem with 1.2.0. Four days ago ftp.redhat.com was running pre2.0.6. Generally Linux's x.even.yy kernels *are* pretty stable because so many people test them to destruction, and Linus doesn't spend millions convincing people that they have to buy Linux95 between midnight and 1:00 on Aug 25th or they're square. >>>>> "Leo" == Leo Glynn <glynn@example.com> writes: Leo> Hi Jim I`ve 1.2.13 Kernel ......... Being waiting for this Leo> release because of load modules . Did you check ppp , I've been bitten once or twice by x.odd.y kernels, but only when I went and got LATEST without checking to see what others were using. JWT's recommendations in particular have been about 100%, maybe 102%. You don't really need to wait for an x.even.y release. Of course, the main reason for getting an x.odd.y module is for a new feature, and these do change over the development cycle introducing the risk of bugs. However, if you make sure that your informants have used that feature successfully the risk is low, not much greater than for the first couple of x.even.y releases (which get used much more heavily by a broader population of systems, so shaking out a lot more bugs). Eg, modules have been stable for me (in fact, have worked increasingly well with each new release) since 1.3.59 when they released the modules-1.3.59 package. Leo> netscape etc. ? Deciding is it worth the move simnce 1.2.13 Is it really that hard to move? Due to various hardware problems I've used a lot of different kernels. So maybe I've forgotten. But it seems to me that it takes about 15 minutes of my time to do the "ftp ftp.cc.gatech.edu; rm linux; tar xvzf linux-NEW-REVISION; mv linux linux-NEW-REVISION; ln -s linux-NEW-REVISION linux; cd linux; make config; make dep; make zImage" and then 15--25 minutes after that I have a new kernel. 10 minutes to re-LILO and viola! So about 1/2 hour of my time. That's a high estimate.... The only thing you need to be careful about is that for development and building stuff from scratch, your current kernel version must match the /usr/src/linux tree. That can even be automated in /etc/rc.d/rc.S (although I don't). You just add a "ln -sf linux-`uname -r` /usr/src/linux" line. (This doesn't work quite right for the pre2 series, since `uname -r` returns "1.99.2" not "pre2.0.2", which isn't obvious from the name of the tar file.) Leo> is pretty stable for me ......... I'll have to upgrade to 2.0. But toriaezu.... Is anybody using kerneld? Netscape works fine with pre2.0.2. ppp works outbound, but I haven't been able to set up a server. That has nothing to do wth Linux AFAIK, the modem won't answer the phone. I think it's because of U. Tsukuba's antique phone system (it's actually an A.G.Bell signature model!), or maybe because I installed HylaFAX and then lost it in the great crash of March '96. Netscape and ppp are not a big test AFAIK. The NET-3 code is pretty stable. What is important to me is that the load avg has gone down although I'm getting more hits on my FTP and HTTP servers, and more telnets from my mail users. I'm not entirely sure this is due to pre2.0.2, since I also doubled my RAM to 32MB at the same time :-) What works for me in pre2.0.2 that doesn't work in 1.2.13: - fdisk now reports the same number of cylinders, sectors, etc on my large hard drives that it formatted them to (something screwy about formatting them under 1.0.9 and examining them under 1.2.13) cfdisk seems more stable in general - modprobe does not get confused about sr_mod and isofs (depmod doesn't order them, in 1.2.13 sr_mod doesn't work reliably, in 1.3.75 the order does matter, in pre2.0.2 it works no matter which order you insmod sr_mod and isofs) - many modules can now actually be inserted and work; under 1.2.13 a lot of drivers did not work as modules for some reason - my Fujitsu 250MB MO drive now can be used to partition and install ext2 filesystems on media without crashing the system; the ext2 fs still does not recognize media change (the directory cache does not get flushed) What has changed: - the /proc fs seems to have changed (although this could be an artifact of the great crash); ps seems discombobulated. You'll need to update your ps I'll let you know if I think of anything else. -- Stephen John Turnbull University of Tsukuba Yaseppochi-Gumi Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp/ Tennodai 1-1-1, Tsukuba, 305 JAPAN turnbull@example.com
- Follow-Ups:
- Re: Linux 2.0.0
- From: jwt@example.com (Jim Tittsler)
- References:
- Re: Linux 2.0.0
- From: Leo Glynn <glynn@example.com>
Home | Main Index | Thread Index
- Prev by Date: Re: Linux 2.0.0
- Next by Date: Re: Linux 2.0.0
- Prev by thread: Re: Linux 2.0.0
- Next by thread: Re: Linux 2.0.0
- Index(es):
Home Page Mailing List Linux and Japan TLUG Members Links