Re: Mirroring - corrections

Pholks, The DBA and I are still trying to figure out exactly what transpired here. It was silly of me to post something like that (but what else is new) before we had all the facts. Neither of us was on location when things happened. In essence - while doing an upgrade - an unfortunate script did a chown root on /dev/*. All but 1 of our engines immediately developed a strange inability to access their disks and aborted. The engine that stayed up was using Bryan's suggestion and had the device driver in a different directory. All 3 engines had their rootdbs mirrored using OS level mirroring vs Informix mirroring. Two of these engines trashed their rootdbs and their mirrors in some manner and had to be restored from tape. It is still way to early to say that a different level of mirroring caused or allowed this problem. That the third dead engine (which also uses OS mirroring) did not trash its rootdbs seems to indicate that the two dead engines may have been in the middle of a transaction and been unable to complete it - thus destroying their integrity. We are still playing detective and may never know exactly what happened. We were already in the middle of a decision to go with Informix mirroring instead of OS mirroring - which is probably why I misunderstood a few things. If I find out anything to enhance this discussion I will let you know. My question is: Does Informix mirroring wait for the first transaction to be successful before trying to duplicate it on the mirror? cheers j. _____________________________________________________________________________ Jack Parker | Hewlett Packard, BSMC Boise, Idaho, USA| "Discover America, jparker@hpbs2561.boi.hp.com | Get lost on a rally." (208) 396-5388 (W) (208) 384-1623 (H) | _____________________________________________________________________________ Any opinions expressed herein are my own and not those of my employers. _____________________________________________________________________________