On Thu, May 04, 2006 at 03:37:53PM +0200, Guido Guenther wrote: > On Thu, May 04, 2006 at 08:35:23AM -0400, Philip R. Auld wrote: > > Hi, > > > > Rumor has it that on Thu, May 04, 2006 at 09:41:20AM +0200 Guido Guenther said: > > > Hi, > > > when using path_checker readsector0 on the above machine the box stalls > > > every 5 seconds when checking the currently nonactive path. Same for > > > path_checker tur. Is there a way to avoid this? I'd consider writing a > > > specific checker but am lacking any documentation. > > > > This thing is a preferred path array. I can see how readsector0 > > is probably the wrong choice of checker. I imagine it causes > > a cache flush from the preferred controller. Odd that the tur > > checker has the same behavior. If you want to code something > Well, can I check which check is really being used, if not I'll hack > that into the daemon next week? I noticed that typos in multipath.conf > are being silently ignored. ...replying to myself: using "path_checker tur" in the defaults section of multipath.conf indeed _does_ seem to work. I've put it to the individual paths before (as shown in several example configs) but this seems to be silently ignored. I'll try to find some time to dig into this next week. Cheers, -- Guido -- dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
