True Story: When your bosses admit his mistake....
Deathmage
Banned Posts: 2,496
So as some of you know I've built out a VMware cluster at my new employer and it has the SAN, the iSCSI fabric, two Dell R720's, and new Dell N2048's switches; the works....
I told them we needed backup power for the servers. Well I got the big fat word we all know and fear "make it work" once they saw the price tag for the cost of adding backup power to the quote.....
Well let put it this way, I managed to jerry rig the sole UPS to keep the SAN alive for 16 minutes with a powerchute cable on it to turn it off instantly as soon as battery backup kicks in. I averaged it take 8 minutes to power off the SAN.
.........Well we had a brown out and then a power outage Sunday and the whole datacenter went offline, and we got the SAN (all the VM's and vCenter database live on the SAN plus the SQL database) down gracefully in 8 minutes and 45 seconds with 6 minutes to spare of the UPS. However we lost one of our servers since the power surge + brown out killed the power supply....
Needless to say, my boss this morning told me to "fix it please" and I remarked with saying "I can't fix this, you should have listened to me" ....he turned and said "yes your right I should, we'll order it right now" .... I said good cause I can replace the power supply and turn the server back on and he was like "really you can do that, yes that I can fix" ...she smiled and lesson learned to my boss..... LISTEN TO YOUR IT STAFF!!!!!
I told them we needed backup power for the servers. Well I got the big fat word we all know and fear "make it work" once they saw the price tag for the cost of adding backup power to the quote.....
Well let put it this way, I managed to jerry rig the sole UPS to keep the SAN alive for 16 minutes with a powerchute cable on it to turn it off instantly as soon as battery backup kicks in. I averaged it take 8 minutes to power off the SAN.
.........Well we had a brown out and then a power outage Sunday and the whole datacenter went offline, and we got the SAN (all the VM's and vCenter database live on the SAN plus the SQL database) down gracefully in 8 minutes and 45 seconds with 6 minutes to spare of the UPS. However we lost one of our servers since the power surge + brown out killed the power supply....
Needless to say, my boss this morning told me to "fix it please" and I remarked with saying "I can't fix this, you should have listened to me" ....he turned and said "yes your right I should, we'll order it right now" .... I said good cause I can replace the power supply and turn the server back on and he was like "really you can do that, yes that I can fix" ...she smiled and lesson learned to my boss..... LISTEN TO YOUR IT STAFF!!!!!
Comments
-
kiki162 Member Posts: 635 ■■■■■□□□□□I have more appreciation for a boss who not only listens, but asks what I can do to help, and actually takes the time to help you when/where he can.
-
IIIMaster Member Posts: 238 ■■■□□□□□□□Maybe she does need to be your boss of she can't follow best practice or know anything about redundant power supplies.
-
dave330i Member Posts: 2,091 ■■■■■■■■■■You have a 2 node cluster for production workload?2018 Certification Goals: Maybe VMware Sales Cert
"Simplify, then add lightness" -Colin Chapman -
Deathmage Banned Posts: 2,496they wouldn't invest in a 3 node cluster, we can run all our server in one host comfortably. the 2nd host is for failover, but when not in failover its @ 50% load balanced. I wish I could get a 3rd host for 33%.
priorities man and budget....baby steps, really..
....was hard enough to get the 30k for the investment in the datacenter. -
Essendon Member Posts: 4,546 ■■■■■■■■■■Good one! What Dave's referring to is to have an n+2 cluster so if you have a host down for maintenance or whatever, you still have a host that can take all the load.
I have general advice for everyone - ALWAYS cover your backside with an email trail. -
Deathmage Banned Posts: 2,496Indeed, we 'can' run all of the servers on the solo server it's just not advisable for best performance since it would have the file server, print server, and SQL on it.
Long-term I'd love to add a 3rd or more hosts. This was really just to get the Private Cloud off the ground and provide proof of concept and the benefits of VMware. A ton rides on my shoulders to provide the level of service that is expected.
Like right now until I DO get a 3rd host, were running one foot in the door one outside the door for the domain controllers so that login's can still occur with a iSCSI connection on the physical server to the SAN (but disabled) so SAN mapping can be sustained in the event of a full-fledged cluster failure.