Re: [cobalt-users] Weirdest memory problem yet...

[Jeff Lasman <jblists@xxxxxxxxxxxxx>]

Glad to see you're around this week, Duncan...

Not only do I need help on this, I've got another question you're
probably best suited to answer; I'll post in a separate post when I'm
done with this reply...

Here are my responses, inline, to your reply...

Duncan Laurie wrote:

> Hey Jeff,
> 
> The swap partitions are setup by `swapon -a' in /etc/rc.d/rc.sysinit,
> which will initialize any swap partition listed in /etc/fstab with type
> "swap" or "sw":
> 
> /dev/hda6             swap      swap    defaults 0 0

Ok.  rc.sysinit has the proper code, and /etc/fstab does as well (I've
taken out some tabular space so the output will work line-by-line in a
post):

/dev/hda1 /        ext2   defaults                         1 1
/dev/hda6 swap     swap   defaults                         0 0
/dev/hda3 /var     ext2   defaults,nosuid                  2 2
/dev/hda4 /home    ext2   defaults,usrquota,grpquota,grpid 3 3
none      /proc    proc   defaults                         0 0
none      /dev/pts devpts gid=5,mode=0620                  0 0

> You can check for swap partitions by running `sfdisk -l /dev/hda' and
> looking for partition type 82, but if you are on a raid system it isn't
> that obvious--by default Cobalt uses /dev/md6 for raid.

This looks wrong:

Disk /dev/hda: 39693 cylinders, 16 heads, 63 sectors/track
Units = cylinders of 516096 bytes, blocks of 1024 bytes, counting from 0

   Device Boot Start     End   #cyls   #blocks   Id  System
/dev/hda1          0+   1524    1525-   768568+  83  Linux
/dev/hda2       1525    1785     261    131544   82  Linux swap
                start: (c,h,s) expected (1023,15,63) found (1023,0,1)
/dev/hda3       1786    2192     407    205128   83  Linux
                start: (c,h,s) expected (1023,15,63) found (1023,0,1)
/dev/hda4       2193   39643   37451  18875304   83  Linux
                start: (c,h,s) expected (1023,15,63) found (1023,0,1)

Any idea why this is happening?

> If the partition type is set to 82, then you can try enabling it with
> `swapon -v /dev/hda6' but if that fails then its possible the swap
> partition is corrupt and can be rebuilt with `mkswap -c /dev/hda6'
> (the -c option checks for bad blocks first)

Is it safe to do this?

> 
> The active (if any) swap partitions are listed in /proc/swaps and can
> also be listed with `swapon -s'.

The table is empty.  Can we do the enable you've mentioned?  Or do we
have to rebuild the entire system?  Or do you have some other idea?

The correct sfdisk table appears to be:

Disk /dev/hda: 25232 cylinders, 16 heads, 63 sectors/track
Units = cylinders of 516096 bytes, blocks of 1024 bytes, counting from 0

   Device Boot Start     End   #cyls   #blocks   Id  System
/dev/hda1          0+   1523    1524-   768095+  83  Linux
/dev/hda2       1524    1845     322    162288    5  Extended
/dev/hda3       1846    2252     407    205128   83  Linux
/dev/hda4       2253   25231   22979  11581416   83  Linux
/dev/hda5       1524+   1584      61-    30743+  83  Linux
/dev/hda6       1585+   1845     261-   131543+  82  Linux swap

Is there any way I can rebuild this on the RaQs in question (there are
now two of them <frown>) without rebuilding them from scratch?

Jeff
-- 
Jeff Lasman <jblists@xxxxxxxxxxxxx>
Linux and Cobalt/Sun/RaQ Consulting
nobaloney.net, P. O. Box 52672, Riverside, CA  92517
voice: +1 909 778-9980  *  fax: +1 909 548-9484