**** BEGIN LOGGING AT Tue Aug 28 02:59:56 2007 Aug 28 03:07:29 crap...something blew up, but I gotta take my fiance to the airport @ like 2am so I need sleep... Thanks for the help rwhitby...I'll pick up and try again tomorrow from here. and let you know how things work out. And I'll get some pictures of the mod. Aug 28 03:07:42 thx Aug 28 05:06:57 I've found an interesting bug that trends across the NSLU2 kernel from 2.6.18 to 2.6.20 (have not yet tried a current one). Seems that even with 400MB swap and mucking about with /proc/sys/vm/overcommit_memory tricks my slug gets confused and thinks it is out of memory. Aug 28 05:07:22 This happens just doing a big tar. Aug 28 05:09:58 I wonder if when the USB drive is busy (which is where the swap partition is) if it times out or something. Aug 28 05:10:18 http://www.pastebin.ca/673138 Aug 28 06:10:51 03bzhou * r6735 10optware/trunk/ (Makefile make/putmail.mk): putmail: added & promoted Aug 28 14:26:12 Good morning all Aug 28 14:27:02 I'm having an interesting issue I've never seen with 3.10-beta Aug 28 14:27:18 I have root and var mounted separately Aug 28 14:28:08 all from a flash drive Aug 28 14:28:22 where sda1 is root, 2 is swap and 3 is var Aug 28 14:28:38 1 and 3 are mounted with 'defaults' Aug 28 14:29:13 but after some interval, aparently less than 12 hours, sda3 suddenly is remounted read-only Aug 28 14:29:29 not via any script I have running, or any other means I can identify Aug 28 14:30:30 are there any other observed cases where filesystems have been inexplicably re-mounted read-only? Aug 28 15:46:04 Anyone have any idea why a system with 400MB of swap and only running a tar extract on a big file from one USB connected disk to another would cause this: http://www.pastebin.ca/673138 Aug 28 15:46:13 NSLU2 Aug 28 15:56:05 VM behavior for recent kernels, as I understand it, apparently will kill programs that lack enough physical RAM regardless of swap space available. Aug 28 16:25:05 mwester: any thoughts on my strange FS remount issue mentioned earlier? Aug 28 16:26:31 I've now explicitly specified the mount options and will leave them in place to see if the issue re-emerges over the next 12 hours, but I haven't been able to identify a prodecure that will precipitate the faulty re-mount, so all I can really do is wait and see Aug 28 16:38:53 mwester: That's really annoying behavior.. :) Aug 28 22:24:46 03bzhou * r6736 10optware/trunk/make/syx.mk: syx: old gcc does not recognize -Wno-strict-aliasing **** ENDING LOGGING AT Wed Aug 29 02:59:56 2007