On Thu, 13 Dec 2007, Richard L. Hamilton wrote:
> Sometimes a large system, despite precautions (or in the absence of them),
> runs out of resources (VM, mainly) to the degree that no useful progress
> is being made: that is, one can't even log in and kill the hogging processes.
>
> (At least on SPARC) the usual workaround would be to break to the boot PROM
> and sync; however, this invariably causes a crash dump to be taken. In the
> common case where dump partition = swap partition, the dump might
> well not be complete anyway, since swap was already full. And one may
> well know by experience what the likely culprits are.
>
> And on a large system, the crash dump can take longer to complete than
> the subsequent reboot!
Sounds like a reasonably simple RFE:
"dumpadm should have option to request skipping dumps"
I.e. like "dumpadm -c none" and/or "dumpadm -d none".
The forced association of "you've got swap we will dump" seems arbitrary.
Note that "swap partition full" doesn't impact the ability to dump, as the
dump will happily overwrite whatever was on there before the dump started.
Shouldn't need PROM modifications that I can see. Just modify dumpsys(),
dump_ioctl() (DIOCGETCONF / DIOCSETCONF) and the dumpadm command. I
estimate around 50 lines of codechange, and takers ?
FrankH.
>
> So it would be really nice to have an option to force a sync and reboot
> _without_ taking a crash dump. The ability to pass the option would require
> boot PROM support (or else some obscure use of an existing command to
> set something the kernel could readily detect); but the ability to optionally
> _not_ take the crash dump even if it otherwise could and would, would
> require kernel support.
>
> Granted, the situation shouldn't come up all that often. But it has been
> known to do so (and one of the culprits tends to be a 3rd-party system
> monitoring application that I won't name; ironic, IMO, that something
> intended to warn of problems instead causes them); and when it does,
> reducing the down-time by the 5-10 minutes that a really long crash
> dump can take, would IMO be quite helpful.
>
> Am I nuts, or is that a generic enough issue that someone else might
> agree that it's interesting?
>
>
> This message posted from opensolaris.org
> _______________________________________________
> opensolaris-discuss mailing list
> [email protected]
>
------------------------------------------------------------------------------
No good can come from selling your freedom, not for all the gold in the world,
for the value of this heavenly gift far exceeds that of any fortune on earth.
------------------------------------------------------------------------------
_______________________________________________
opensolaris-discuss mailing list
[email protected]
