vBulletin Search Engine Optimization
| |||||||
| Register | FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
| ||||
| Hi. We have a 2 x 44P and 2 x SSA enclosures in our platform. Each SSA is looped around both of the 44Ps, so that's pretty simple. (44P 1 SSA card port A1 -> SSA enclosure 1 -> 44P 2 SSA card port A1; 44P 2 SSA card port A2 -> 44P 1 SSA card port A2. And the same for the SSA cards' B ports using the second SSA enclosure.) This platform is running 24x7 but we now are about to upgrade the disks in each SSA enclosure from 4x9GB to 1x9GB + 4x36GB. The disks are mirrored from one enclosure to the other. We have another identical platform that we are testing the changes on. The time taken to set up the disks with mkvg, mklv, mklvcopy, and crfs is about 30 minutes, but the syncvg command takes over 5 hours. Our current plan to minimise operational down-time is to use mksysb on the operational machines, install that on the test machines, set up the disks, run mksysb on the test mchines, and install that back on the operational machines at the same time as swapping the disks. This should keep operational down-time to less than an hour. We don't plan to swap the enclosures themselves due to their bulk and the fact that the operational and test platforms are in different cities. Does this sound feasible and sensible? Thanks for your time. Jeffrey |
| ||||
| "Jeffrey Ross" <jeffrey.rossATairways.co.nz@no.spam> wrote in news:40984ac2$1@news.iconz.co.nz: > Hi. > We have a 2 x 44P and 2 x SSA enclosures in our platform. Each SSA is > looped around both of the 44Ps, so that's pretty simple. (44P 1 SSA > card port A1 -> SSA enclosure 1 -> 44P 2 SSA card port A1; 44P 2 SSA > card port A2 -> 44P 1 SSA card port A2. And the same for the SSA > cards' B ports using the second SSA enclosure.) > This platform is running 24x7 but we now are about to upgrade the > disks in each SSA enclosure from 4x9GB to 1x9GB + 4x36GB. The disks > are mirrored from one enclosure to the other. > ... > Does this sound feasible and sensible? A bit faster would be to unmirror one enclosure, delete disks from the ODM, replace old disks with new, cfgmgr, mirror the VG again, and repeat for the first enclosure. This could all be done online. If you are using hacmp, you need to export/import the volumegroup on the standby system. Unless of course you use concurrent mode, in which case the above solution is not useful at all :-) cheers, Maarten Kreuger |