On Thu, Jan 26, 2017 at 2:12 PM, Olaf Marzocchi <li...@marzocchi.net> wrote:
> But dumps can also be saved as files on a normal dataset, right? provided > enough space is left for them. > No. The dump is a two-stage process. When the system panics, it simply drops memory into the dump volume. (Traditionally, it used to use the swap partition.) Then, when the system is back up, you save that dump into regular files for subsequent analysis. If you're really tight for space, and aren't worried about debugging a panic, then disabling dumps entirely (dupadm -d none) might be appropriate in this case. (As an aside, I note that current OmniOS LTS - r151014 - doesn't understand dumpadm -e, which is a shame.) > Olaf > > > > Il 26 gennaio 2017 12:38:27 CET, v...@bb-c.de ha scritto: >> >> NAME USED AVAIL REFER MOUNTPOINT >>> rpool/dump 41.5G 9.15G 41.5G - >>> rpool/swap 4.13G 13.0G 276M - >>> >> >> The "dump" volume is much too big. Do a >> >> dumpadm -e >> >> This will print the "estimated" dump size. Then add a bit, and >> set the new dump volume size with: >> >> zfs set volsize=<new size> rpool/dump >> >> For example, on my OmniOS file server: >> >> # dumpadm -e >> Estimated dump size: 4.63G >> >> # zfs set volsize=6G rpool/dump >> >> # zfs list rpool/dump >> NAME USED AVAIL REFER MOUNTPOINT >> rpool/dump 6.00G 24.1G 6.00G - >> >> I did not changed anything during instalation proccess, I've just >>> accepted all defaults >>> >> >> Yes. The "traditional" installation usually sizes dump too big. >> That is why I asked. :-) >> >> >> Hope this helps -- Volker >> >> > _______________________________________________ > OmniOS-discuss mailing list > OmniOS-discuss@lists.omniti.com > http://lists.omniti.com/mailman/listinfo/omnios-discuss > > -- -Peter Tribble http://www.petertribble.co.uk/ - http://ptribble.blogspot.com/
_______________________________________________ OmniOS-discuss mailing list OmniOS-discuss@lists.omniti.com http://lists.omniti.com/mailman/listinfo/omnios-discuss
