[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

i586 issues

Having not really done a fresh install on any of my 586 hardware lately,
(yum upgrade works for me, most of the time...) I tried both test 1 and
2 rescue cd, I like the idea of using stage2 from the cd for a text/http
install. While both boot the kernel, and initrd runs, once stage2 is
mounted and anaconda starts, the install just shuts down, with 256M of
ram installed. The same cdrom and cdrom drive when placed in an i686
machine with less ram, run though to completion. Sorry, I didn't think
about using remote syslog at the time. Is this the CMOV rpm/glib issue?
I'm thinking that busybox and/or the glibc libs might be compiled for
i686 and that is what is causing this issue.  any thought? (don't say
scrape the junk ;-))  Should I file a Bug report?

Not to be discouraged, I slapped a hard drive into a usb adapter, used a
faster machine, but got caught on a uname issue. Playing around with a
kickstart file's +/- packages didn't fix it. Remembered seeing a
targetarch option in flags I gave that a shot, but didn't work either. I
was thinking of jumping distros at that point, after I've been using RH
since 6.1... After looking at the code in iutil, I hacked up yuminstall
with the attached patch. This worked while using rootpath, I'll give it
further testing later. Which leads me to wonder if the issue in bz#
206881 is going to crop up again. Should I file a Bug report?


--- yuminstall.py.orig	2007-10-08 16:03:59.000000000 -0500
+++ yuminstall.py	2007-10-08 21:07:06.000000000 -0500
@@ -900,6 +900,20 @@
+    def fake586(self):
+	    ### Jerry's hack ###
+        myarch = rhpl.arch.canonArch
+        log.info("Detected %s as arch", myarch)
+        # now allow an override with targetarch=i586 on the command line 
+        if flags.targetarch == "i586":
+            myarch2 = flags.targetarch
+            log.info("Detected cmdline targetarch %s as arch", myarch2)
+        # now make the current install believe it, too
+            for pkg586 in ['glibc.i386', 'kernel.i586']:
+                self.selectPackage(pkg586)
+                log.info("Selected %s for installation", pkg586)
+        ### end ### 
     def doPostSelection(self, anaconda):
         # Only solve dependencies on the way through the installer, not the way back.
         if anaconda.dir == DISPATCH_BACK:
@@ -909,6 +923,7 @@
         self.ayum.dsCallback = dscb
         # do some sanity checks for kernel and bootloader
+        self.fake586()
         self.selectFSPackages(anaconda.id.fsset, anaconda.id.diskset)

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]