My Monday Wasn't Supposed to be Like This
May. 12th, 2008 10:37 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
bjarvisI left work at my usual time this afternoon but the heavy rains slowed commuter traffic considerably. Thanks to some judiciously chosen back streets --dodging the flooded areas of town-- I still made it home without being too late.
On a whim, I powered up my office laptop and logged in to see if there was anything particular in e-mail since I left work. There was. Hardware work that was originally scheduled for Tuesday night had been shifted to Monday. Not a big deal, except that the hardware to be installed was locked in my cube and I had the only key. Damn!
The return commute to the office would have been pure hell at that moment so I had a brief dinner, took a nap and then, while![[livejournal.com profile]](https://www.dreamwidth.org/img/external/lj-userinfo.gif)
kent4str and cuyahogarvr went to justetthon's C1 dance, I drove the other way towards Frederick, MD.
The night shift had arrived early so I was able to hand them the hardware directly and discuss the task list with them. We came up with a small short-cut for this particular project which will save some time later this week too.
The project is to migrate data from the Sun Microsystems D1000 local disk arrays attached to Sun E420R servers to our SAN. Because of the lack of available PCI slots in the E420R servers, we were going to remove one SCSI card and install one SAN card tonight, pass the world-wide numbers to the SAN team tomorrow so they could do the SAN fabric zoning Tuesday night, then reboot & mirror the Veritas volumes of data Wednesday night, run integrity tests Thursday night and replace the remaining SCSI card with a second SAN card Friday night, breaking the mirrors, adding the new card to the SAN zoning and removing the local disk array entirely. This would be done on five interrelated machines per week.
The lead night guy is a former Sun Microsystems field engineer and he mentioned that the SCSI cards in question were dual channel: by rearranging the SCSI cables, we could keep dual SCSI paths to the local disk array, even with one SCSI card removed. Better yet, we could remove the frame buffer card entirely to make space for both SAN cards to be installed tonight so the zoning work would only have to be done once. This will make Friday's work much easier, a valuable thing for us since our Friday night shift is usually very heavily loaded.
I'll just have to keep my pager on my nightstand all night, just in case. If this works as projected, we'll apply it to the other SAN migrations scheduled this month and next.
I drove directly to the C1 dance from the office, arriving in time for the last half-hour of dancing.kent4str tapped out so I could dance with cuyahogarvr. Both cuyahogarvr and meinfs were dancing easily and smoothly. While our squares occasionally had issues, the problems were just as often by the more experienced dancers as by our newest ones. Declare victory, I say!
On a whim, I powered up my office laptop and logged in to see if there was anything particular in e-mail since I left work. There was. Hardware work that was originally scheduled for Tuesday night had been shifted to Monday. Not a big deal, except that the hardware to be installed was locked in my cube and I had the only key. Damn!
The return commute to the office would have been pure hell at that moment so I had a brief dinner, took a nap and then, while
kent4str and cuyahogarvr went to justetthon's C1 dance, I drove the other way towards Frederick, MD.The night shift had arrived early so I was able to hand them the hardware directly and discuss the task list with them. We came up with a small short-cut for this particular project which will save some time later this week too.
The project is to migrate data from the Sun Microsystems D1000 local disk arrays attached to Sun E420R servers to our SAN. Because of the lack of available PCI slots in the E420R servers, we were going to remove one SCSI card and install one SAN card tonight, pass the world-wide numbers to the SAN team tomorrow so they could do the SAN fabric zoning Tuesday night, then reboot & mirror the Veritas volumes of data Wednesday night, run integrity tests Thursday night and replace the remaining SCSI card with a second SAN card Friday night, breaking the mirrors, adding the new card to the SAN zoning and removing the local disk array entirely. This would be done on five interrelated machines per week.
The lead night guy is a former Sun Microsystems field engineer and he mentioned that the SCSI cards in question were dual channel: by rearranging the SCSI cables, we could keep dual SCSI paths to the local disk array, even with one SCSI card removed. Better yet, we could remove the frame buffer card entirely to make space for both SAN cards to be installed tonight so the zoning work would only have to be done once. This will make Friday's work much easier, a valuable thing for us since our Friday night shift is usually very heavily loaded.
I'll just have to keep my pager on my nightstand all night, just in case. If this works as projected, we'll apply it to the other SAN migrations scheduled this month and next.
I drove directly to the C1 dance from the office, arriving in time for the last half-hour of dancing.
kent4str tapped out so I could dance with cuyahogarvr. Both cuyahogarvr and meinfs were dancing easily and smoothly. While our squares occasionally had issues, the problems were just as often by the more experienced dancers as by our newest ones. Declare victory, I say!
