Re: VMS732_SHADOWING V4.0, mount verification, solid-state disks
From: Phillip Helbig---remove CLOTHES to reply (helbig_at_astro.multiCLOTHESvax.de)
Date: 07/05/05
- Next message: Phillip Helbig---remove CLOTHES to reply: "Re: [Announce] FreeVMS 0.1.7"
- Previous message: Phillip Helbig---remove CLOTHES to reply: "Re: VMS732_SHADOWING V4.0, mount verification, solid-state disks"
- In reply to: Ed Wilts: "Re: VMS732_SHADOWING V4.0, mount verification, solid-state disks"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Tue, 5 Jul 2005 19:59:40 +0000 (UTC)
In article <1120576557.960406.193230@o13g2000cwo.googlegroups.com>, "Ed
Wilts" <ewilts@ewilts.org> writes:
> Phillip Helbig---remove CLOTHES to reply wrote:
> > after which I noticed the following problem: After mounting the shadow
> > set, which immediately goes into a merge state (probably unavoidable due
> > to the page and swap files, but since it's solid-state, it is quick),
>
> The HBMM kit may help with this. At least you'll do a mini-merge
> instead of a full merge.
Under normal circumstances, a full merge is just a minute or so. Even
if it were quicker, the problem would presumably still remain.
> What's the time frame on the mount verification messages?
Several per second. It might even be that the bottleneck is the console
speed, i.e. there were OPCOM messages at the rate of several per second,
but the actual messages might have been even more frequent.
> Do they go
> in an out on a regular basis (like every minute) or is it continuous?
Continuous and regular, as far as I can tell.
> If the former, I know of at least 2 customers that are experiencing
> this on a consistent and reproducable basis (I'm one of them and the
> call has been escalated to Engineering).
Interesting. RZ or EZ disks? 7.3-2? After installation of the latest
shadowing patch? If the rate is slower than mine, is it slower by the
ratio of the speed of an EZ to an RZ disk? :-)
> What happens if you drop one member of the shadow set and do a full
> copy instead of a merge?
Haven't tried it. Basically, I would need more time than I have to do a
full test. It appears the only way to stop things is to power-cycle the
machine, which I don't like to do often. Also, with a DEC 3000 LX with
48 MB and 7.3-2, booting takes a long time. I can kill the process
doing the mount from another node in the cluster (by doing a STOP/ID in
a loop in DCL to catch it when it is not suspended), but even after that
things are hung up so that I have to reboot.
- Next message: Phillip Helbig---remove CLOTHES to reply: "Re: [Announce] FreeVMS 0.1.7"
- Previous message: Phillip Helbig---remove CLOTHES to reply: "Re: VMS732_SHADOWING V4.0, mount verification, solid-state disks"
- In reply to: Ed Wilts: "Re: VMS732_SHADOWING V4.0, mount verification, solid-state disks"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
