I had a good back. Until some bitch failed to yield and I T-boned her.
Entered on 11/23/2011 at 15:34:56 PST (GMT-0800) by Ryan Ch.: Hi Erik,
This is correct, though the hidden folder cannot be accessed while as Suggested Contacts is also used for autocomplete and can be accessed.
Entered on 11/23/2011 at 15:01:21 PST (GMT-0800) by PLR_3104_48274: Autocomplete and Suggested Contacts are two separate entities.
Entered on 11/23/2011 at 14:29:54 PST (GMT-0800) by Ryan Ch.: Hello Erik,
In Exchange 2010 these are stored in the "Suggested Contacts" folder which you can see within Outlook. You can backup that folder if you are worried about losing the data.
What I know for certain is that I am overweight. I decided a few weeks ago to get a head start on my New Year's resolution; so now I do at least 500 kcals of work on the elliptical machine. every. damn. day.
(okay I have skipped a day here and there, but I feel guilty when I do)
I decided I had to take a page from Limoncelli's book and just precompile the decision to the simplest thing possible. Any complexity, any "this day but not that day" stuff just lets me play mental games and outsmart myself into not exercising
bingo. and the elliptical machine is pretty forgiving - you can work hard at it, or not ... but the kcal thingy does keep going up as long as you keep moving the pedals
(at different rates obviously)
I won't lie - it's still boring, and I still procrastinate. Sometimes I wait till 10 at night. Podcasts and audiobooks help with the boredom.
@Holocryptic The Outlook autocomplete stuff is a pain in the bum. It became absolutely unmanageable with 2010. What a crazy idea to have a "hidden folder" no one is able to access or back up properly. And the hosters make it worse by not providing interfaces for a full mailbox backup, although it would be rather easy to set up.
@quux I think I'll have to get something like that, yeah
They should just incorporate the suggested contacts with the auto complete, and quit fucking around with it. It's so many hours of stupid mindless work trying to fix it when it goes wrong
and it always goes wrong
Its the whole cache thing that gets screwed up.
But now users are dependent on it. Like public folders. So MSFT says "We're doing away with PF, but we'll still have backwards compatibility and give you a shitty way of administering it"
I know many users who do not have contact lists or address books anymore but manage everything through autocomplete. That is at least until it breaks for the second or third time.
And hey: who needs public folders when there is sharepoint? har har
The software is there to empower the user - everybody knows that empowerment is a good thing. Why shouldn't it also empower users to break as much as they possibly can?
@Holocryptic we are using OTRS as the ticketing system - it is able to render HTML mails but of course no Word documents. Our users keep sending us mails with screenshots in Word attachments without any hint in mail body or subject what it is about - so we've lost track one time or the other.
I have been asked to truncate a large table in MS SQL Server 2008. The data is not needed but might be needed once every two years. It will NEVER have to be changed, only viewed.
The question is, since I don't need the data on a day-to-day basis, what do I do with it to protect and back it up?
...
We're moving offices in feb and I'm figuring out how much cooling our new comms room will need. Would it be ok if I posted a "Check my working" question, or would that invite too much discussion?
what was the conclusion about non-VMWare hypervisors? We're quite interested in what MS are promising for windows server 8's version of hyperv. That might well be "good enough" for us
I'm sufficiently interested in HV, Xen, KVM, OVM etc. to keep an eye on them but their cost benefits don't outweigh the flexibility and depth of toolset that vSphere does yet, but I'm still watching that segment closely
The nice features of our HV clusters haven't cost the earth (as we're educational) and we don't have a need for most of the nice features VMWare offers (Despite my desire we use it)
@ITHedgeHog we're educational too but we need vmotion and storage vmotion
If Microsoft can get their windows 8 version of hyperv to where vmware are now (or even where they were with 4.1) then we can probably switch to hyperv
@pfo I agree... Microsoft and VMware give the hypervisor away and sell you tools to manage them
@pfo intel do that already - their chips (and AMD's) do 99% of the heavy lifting of the actual virtualisation these days anyway - obviously they don't provide a means to do the more advanced functionality and that's where vmware etc come in, as does the varience in offering
@pfo orchestrator's a bit rubbish but it's free with VC, cloud director is miles better but spendy - I think SRM's very well priced tbh
No, it'll be debian on the OS, we're centralising some MySQL analytics work on it. It will be the first iLO3, though. Our c3000 is on iLO2. Looking forward to taking a look at the new one.
If you ever feel the need to clear the screen in chat (perhaps to make it sfw then just run this javascript:void($('#chat').empty()) in you browser's javascript console
Oh, if you're still at 15, you might not have the lvl 3 unrelenting force shout, yet. I recommend getting that. It's basically force push. Hilarious when chucking bandits off mountains.
one of the managers screamed at support, saying that mails from other co-workers didnt arrive to him.. we checked the exchange logs, and could clearly see that it was delivered to his inbox
turns out that he has 367543146351 different outlook rules to automatically place mails into separate folders, and he claims to have full control of them, and that the problem must lie elseware
so thats pretty much where we are right now.. he's confident that there is something wrong with the server.
Does anyone know of any good sites where I could get an up to date analysis of enterprise class SSD performance? I've been under the assumption for years that they're great for random reads, ok for sequential reads (compared to spinning disk), and the same or slightly worse than spinning disk for writes.
I've been told by 4 vendors over the last week that their disks are fast at writing
@Holocryptic The issue is that the benchmarks are on the OEM SSDs, where I'm looking to see how they perform when they're used by, for example, HDS or EMC in an array
@WilliamHilsum Hey Wil. SF.Chat tends to pick up around 7pm, although there's less work/sensible discussion and it tends to lean towards memes. And ponies. Lots of ponies.
@Holocryptic The only independent benchmarks are SPC, and those are for the whole machine. I'm more looking for something that states "SSDs come in two flavors- those that read quickly and those that write quickly" or something like it.
I could be wrong but I would think that reads and writes would be very similar. You could ask @chopper3 hes got pretty good insight on large scale hardware things
@Basil I don't have any numbers, but EMC claims that you would save money by buying a SSD array from them compared to the amount of disks you'd need for the same sequential iops
let me check our VNX, maybe I can fish out the real model number
we went the fast cache only route.. atleast for now
fast cache + 10k + 7,2k
im a bit unhappy about the tiering.. it puts all new data on the slowest media, and then moves it up after a while.. I think it should have been the other way around
we had to pin certain luns to the 10k disks to get out of trouble when doing large new datasets etc
looks like they've hidden the vendor.. ill put up a screenshot so that you can dig for whatever you want
ok it's samsung.. "SAMSUNG SS160510 CLAR100"
"In October 2010 - Samsung said it is shipping 200GB 3.5" SATA SLC SSDs to EMC. Sequential R/W speeds are 260MB/s and 245MB/s respectively. R/W IOPS are 47,000 and 29,000. The new Samsung SSDs have an 'end-to-end data integrity' function and encryption."
@Holocryptic It is related, but I'm not 100% familiar with the rules about what is "related" and what is a "new question".. Also, I'd probably have to rewrite the title, rewrite the question itself, etc.
np mate.. good luck. Just keep in mind that FAST Cache doesnt help much if you got 600GB of data thats being constantly written/read (unless you have 12 ssd's wich totals in 600GB of cache).. you'd be better off with ssd's in tiering or as a separate array
@robinhoode No, it would not
Your original question basically states "My car broke down. What is wrong with it?"
if you updated it, it would be "My Car broke down, and there is oil dripping from the cylinder. What could the reason be?"
which one do you think is easier to help you with?
@pauska You're absolutely correct, but I feel that changing the question at this point, given that it's already got 4 different answers, will not attract much attention.
@Holocryptic Yeah, that's what I'm saying. My previous question was about "potential intruders". This is more about diagnosing why a particular process was called. The overlap isn't very large (to me at least).
@Chopper3 I just have to get a proof of concept working with HAproxy both load balancing and fault tolerance against two web servers - I think I am finally there... it was an error extracting the TAR, so, it explains why it was failing the build!
and... learning Linux, I am getting there - just slower than I would like (I'm not exactly a noob, but, been trying to master it for ages, this is the excuse I need!)
if you have a fixed, pre-known LUN that you know needs either amazing random read on or pretty good random write then they're right, a pair of SSDs will be cheaper (depending on discount level) than a bunch of 15k's - also bear in mind the disk-striping aspects of your array, many won't stripe a LUN over more than say 8 disks
if the LUN's smaller than the available space on that 8 disk block - meaning that if you have say 200GB of high-write data on an array with 128 x 300GB 15k's your data may still only ever be served by just 8 of them - understand? Now if it's being used to dynamically move high-traffic blocks/clusters from slower disks this is ALWAYS a good thing as long as you've got enough ssd to cover this high-traffic data anyway.
if you have a fixed, pre-known LUN that you need either amazing random read on or pretty good random write then they're right, a pair of SSDs will be cheaper (depending on discount level) that a bunch of 15k's - also bear in mind the disk-striping aspects of your array, many won't stripe a LUN over more than say 8 disks if the LUN's smaller than the available space on that 8 disk block
- meaning that if you have say 200GB of high-write data on an array with 128 x 300GB 15k's your data may still only ever be served by just 8 of them - understand? Now if it's being used to dynamically move high-traffic blocks/clusters from slower disks this is ALWAYS a good thing as long as you've got enough ssd to cover this high-traffic data anyway.
so yes, I agree with a lot of what you put there, but the only way we'd need SSDs for our OLTP workload is with something that automatically moves data between tiers. That said:
@pauska That sense of arrogance and profit-taking is what's caused them to lose some traction chez nous. Even the storage team lead who worked there for 5 years
the NVLOG device in a netapp can perfectly fine absorb random write IO and pass sequential writes to RAID-DP which can do sequential writes quite nicely if it has enough free space
I have been asked to truncate a large table in MS SQL Server 2008. The data is not needed but might be needed once every two years. It will NEVER have to be changed, only viewed.
The question is, since I don't need the data on a day-to-day basis, what do I do with it to protect and back it up?
...
