|
Docjowles posted:You already got some good answers, but another decent tool I ran across recently is Microsoft's DiskSpd (formerly SQLIO). This is what Veeam recommends to test performance for their app.
|
# ? Aug 31, 2016 14:20 |
|
|
# ? Apr 17, 2024 19:06 |
|
I think the "cifs shares" command will also let you change permissions.
|
# ? Sep 1, 2016 23:56 |
|
Nothing says ENTERPRISE STORAGE quite like an crusty emeritus showing up to the servicedesk with a box full of floppies.
|
# ? Sep 2, 2016 15:36 |
|
I just...I just give up. I've been working through a new datacenter deployment/migration, and we are all ready to go for the migration phase except for the new SAN install. We got our new SC8000 pair installed in the colo this week, and our last step was to configure replication from our current production SC030 compellent to the new one for the migration process. We also purchased another pair of SC8000s at the same time for DR with a couple of disk shelves. The SC30 SCOS is too old and the SC8000 SCOS is too new; there's a SCSI I/O card driver mismatch which breaks replication, which we discovered after an hour of troubleshooting and finally calling Copilot support - we had the answer in 10 minutes. If we backrev the SC8000s enough for the driver issue, we would lose access to the SSD tier until we are completely migrated and can upgrade the controllers. Unacceptable, as I have a really dumb ERP sql system that I need to throw more IOPS at as soon as I can migrate it. We can't upgrade the SC30 to a high enough rev because they are 32-Bit only, latest versions of SCOS are 64-bit. So my deployment consultant believes that we can set up the new SC8000s for DR in the colo (which we would need to do anyways for initial replication seeding), and set up a two-step replication process, from our current prod, to new DR SC8000s over the 100mbit mpls, to direct to the new production SC8000s over 10GBE. However, we are missing a set of compellent 10gbe cards, because we were going to take the existing ones in current production down to the new DR, after we did the datacenter migration and had it decomissioned. I talked to our dell rep to see what options we have, I was hinting pretty strongly that he should find me some loaners for a few months. If not, get me a god damned quote ASAP so I can call my boss on her day off to approve it.
|
# ? Sep 2, 2016 19:33 |
|
I think the important lesson here is to not buy Compellent.
|
# ? Sep 2, 2016 19:44 |
|
Did your VAR know what you wanted to do with the new boxes and sold them to you with that in mind? Make it their problem.
|
# ? Sep 2, 2016 19:53 |
|
Thanks Ants posted:Did your VAR know what you wanted to do with the new boxes and sold them to you with that in mind? Make it their problem. I'm commissioning our new netapp and JFC raid-dp writes are so loving slow. It's going to take 5 days to replicate 8(
|
# ? Sep 2, 2016 20:25 |
|
Everything was vetted and we were good for install, the problem is because of the drive mismatch dell introduced with the latest SCOS. There isn't even a bulletin out there, but once we described to the scenario to copilot support he instantly knew what the problem was. That was our last step to get it 100%.
|
# ? Sep 2, 2016 20:37 |
|
I still think that isn't your problem to sort out, and they need to be working with Dell rather than just having you buy more poo poo. Obviously saying "not your problem" doesn't solve the immediate issue, but it sounds like your hands are tied anyway.
|
# ? Sep 2, 2016 20:49 |
|
evil_bunnY posted:Holy poo poo this. How much are you replicating? Depending on the controller I've seen snapmirror max out a 10Gb link.
|
# ? Sep 2, 2016 21:02 |
|
NippleFloss posted:How much are you replicating? Depending on the controller I've seen snapmirror max out a 10Gb link.
|
# ? Sep 2, 2016 21:05 |
|
Ignore this, I am an idiot
|
# ? Sep 2, 2016 21:11 |
|
evil_bunnY posted:~40TB to a 2552 on 10GBE but the destination is wide rear end sata aggregates, so it just pegs the cpu. I'm pretty sure it's the writes since the read side will happily do multiples of the speed I'm getting now (65MBps). It's the CPU specifically. The 2552 has a dual core CPU that's like six years old. They didn't change them at all between the 24xx and 25xx controllers. It should still be somewhat faster than that though, if the destination aggregate is empty and there's no other background work going on on the destination.
|
# ? Sep 2, 2016 21:26 |
|
NippleFloss posted:It's the CPU specifically. The 2552 has a dual core CPU that's like six years old. They didn't change them at all between the 24xx and 25xx controllers.
|
# ? Sep 2, 2016 21:29 |
|
The 24xx and 25xx are really branch office systems. SnapMirror is also very CPU time sensitive, doubly so if you have any other workloads sitting on them. They do have 4 cores effective. Wrar fucked around with this message at 17:02 on Sep 4, 2016 |
# ? Sep 4, 2016 17:00 |
|
I've been out of the loop on storage stuff for a while and we're needing to expand ours at work - can I just get a quick sanity check on my plan below? I just want to make sure I'm not doing anything colossally stupid. We have an older Stornext SAN that we can't afford to keep paying the maintenance and upgrades on, and we need to increase the available space. Since our current setup is based on commodity hardware (Promise vtrak chassis'), my plan is to dissolve the SAN as we need know it and attach the Promise arrays to a Windows server 2012, essentially turning it into an 10g SMB3 NAS (we have no need for a direct-attach FC SAN anymore). What we'd end up with is 3 Promise E830f's fully populated with 6TB SAS disks, connected to our server via 8g FC. I was planning on configuring each of them as a single large RAID5 (or 6) array and having 3 separate storage pools. I've heard of people breaking them up into smaller RAIDs on the chassis and then raiding those together in Windows, but I'd like to keep this is a simple as possible and avoid any software raid (plus I'm not sure it's really necessary from a performance standpoint). Is this stupid? Will I run into any headaches with these large (~130T) NTFS volumes?
|
# ? Sep 26, 2016 22:11 |
|
A 24 disk RAID6 fills me with dread
|
# ? Sep 26, 2016 22:16 |
|
Thanks Ants posted:A 24 disk RAID6 fills me with dread Because of the possibility of multiple disk failures or what? What configuration would you recommend, then? (i'd have at least 1 hot spare per chassis fwiw)
|
# ? Sep 26, 2016 22:33 |
|
n.. posted:Because of the possibility of multiple disk failures or what? What configuration would you recommend, then? (i'd have at least 1 hot spare per chassis fwiw)
|
# ? Sep 26, 2016 22:38 |
|
Vulture Culture posted:Here's the hypothetical that drives the whole rest of this storage layout: what do you think the rebuild time is on a 6 TB, 7200 RPM disk, assuming zero utilization of the array in the meantime? Hm, I'm not sure but it's probably a longass time, 12+ hours if I had to guess. I see what you're getting at here. So probably best to sacrifice some space and go with RAID60 or something.
|
# ? Sep 26, 2016 22:46 |
|
n.. posted:Hm, I'm not sure but it's probably a longass time, 12+ hours if I had to guess. I see what you're getting at here. So probably best to sacrifice some space and go with RAID60 or something. Days with application IO happening concurrently.
|
# ? Sep 27, 2016 01:38 |
|
NippleFloss posted:Days with application IO happening concurrently.
|
# ? Sep 27, 2016 12:43 |
|
So am I correct in assuming the best way to mitigate that concern, assuming I'm stuck with the 6TB drives in 24-bay enclosures for budgetary reasons, is to use RAID60? Say, three 8-disk RAID6's per enclosure? The rebuild times would still be just as long, but at least it would only affect a third of the drives at a time instead of all of them. (no application data on these, pure storage and mostly sequential IO)
|
# ? Sep 27, 2016 17:49 |
|
n.. posted:So am I correct in assuming the best way to mitigate that concern, assuming I'm stuck with the 6TB drives in 24-bay enclosures for budgetary reasons, is to use RAID60? Say, three 8-disk RAID6's per enclosure? The rebuild times would still be just as long, but at least it would only affect a third of the drives at a time instead of all of them. (no application data on these, pure storage and mostly sequential IO)
|
# ? Sep 28, 2016 04:07 |
|
Need some NetAPP advice. I just inherited a Dual Head FAS3240 with no login/password. I was able to recover the support account used and download OnCommand. It's running 8.2.4 in 7-mode. I *short-term* have console access while I'm on site. And just like the rest of my luck, it's running production critical storage. Is there ANY way I can reset the user/password without taking it offline? Oh and it literally fell out of the maintenance the day the former admin left. It was one of the very few parting bits of info he passed on. We're working to renew it/figure out what to do with it.
the spyder fucked around with this message at 23:06 on Oct 4, 2016 |
# ? Oct 4, 2016 22:34 |
|
the spyder posted:Need some NetAPP advice. I just inherited a Dual Head FAS3240 with no login/password. I was able to recover the support account used and download OnCommand. It's running 8.2.4 in 7-mode. I *short-term* have console access while I'm on site. And just like the rest of my luck, it's running production critical storage. Is there ANY way I can reset the user/password without taking it offline? Oh and it literally fell out of the maintenance the day the former admin left. It was one of the very few parting bits of info he passed on. We're working to renew it/figure out what to do with it. Edit: I can't actually think of a way to do this without being able to log in to do a giveback. If auto-giveback was configured it would work, but you wouldn't know that. Have you tried the usual suspects for default passwords? Was access for domain accounts never granted? Are there any service accounts for things like snapdrive or VSC that connect to the NetApp? YOLOsubmarine fucked around with this message at 23:56 on Oct 4, 2016 |
# ? Oct 4, 2016 23:48 |
|
the spyder posted:Need some NetAPP advice. I just inherited a Dual Head FAS3240 with no login/password. I was able to recover the support account used and download OnCommand. It's running 8.2.4 in 7-mode. I *short-term* have console access while I'm on site. And just like the rest of my luck, it's running production critical storage. Is there ANY way I can reset the user/password without taking it offline? Oh and it literally fell out of the maintenance the day the former admin left. It was one of the very few parting bits of info he passed on. We're working to renew it/figure out what to do with it. Wow, between this and your post in the Cisco thread it sounds like you've gotten into quite the situation. Good luck mate.
|
# ? Oct 5, 2016 00:17 |
|
Since you seem to be hosed multiple ways, let me ask this: how big is the 3240? If it is out of support, is it worth just replacing it with something else and hoping you can keep it online until that is complete?
|
# ? Oct 5, 2016 03:12 |
|
I've tried every combo of logins, passwords, domain admin accounts, ect. It appears to be on the domain, as it requires a domain user account to authenticate to the shares, but I'm wondering for how long. The domain admin account was also changed, which was most likely used as the credentials for auth in OnTAP. The only note he left was to use na_admin, which 1) was discontinued in 8.1 and 2) does not work as a login in combo with any of the provided passwords. It's a full-rack of 2TB and 3TB shelves. The worst part is we've only had it 6 months and inherited it from yet another site. We were arguing wether or not to keep it due to the cost of the maintenance renewal. I literally have no idea why it was put into production, yet alone for the rather measly 45TB of data it's hosting when we have a ~9 month old Isilon that has plenty of capacity/performance. Thankfully it appears to only be hosting SMB shares. I'm going to see what we need to do to migrate the data back to the Isilon. Production staff may hate the downtime, but I can not keep this system in production while there's a chance it could drop auth and loose access to the data. I just can't even process how many other things are setup just like this. 70% of the equipment needs the scorch and burn to clean ground treatment- if we can get the business to sign off. *Edit* Does anyone know how often OnTAP polls/uses the authentication credentials? *2nd Edit* Just checked the AD servers. Log has a handful of errors relating to being unable to authenticate the account referenced in the security database. Wonder how long until logins to the shares drop... the spyder fucked around with this message at 05:32 on Oct 5, 2016 |
# ? Oct 5, 2016 04:10 |
|
I need help with choosing an enterprise NAS for my video editing team! We're looking for a way to backup and easily access our video project files (we do video production for a large company). Right now, we store everything on daisy-chained external HDs. And it's terri-bad. We need something better than external HDs. We also need something that can hook up to our office's Internet network. At home, I use the Synology DS214 Play for backing up and accessing my personal project files. It's a great setup that I've been really happy with. Is there anything similar to this that'd be better suited for a larger office for files that need to be accessed by 3-4 employees at the same time? And should we be going for DAS, or NAS (or even SAN)? Edited to add our requirements:
melon cat fucked around with this message at 15:55 on Oct 5, 2016 |
# ? Oct 5, 2016 15:20 |
|
Figure out your capacity and performance requirements first.
|
# ? Oct 5, 2016 15:30 |
|
melon cat posted:I need help with choosing an enterprise NAS for my video editing team! Video is a special usage case of these types of devices. There are a LOT of things to consider: 1) What kind of network do you have in place? A cheap or lovely gigabit network will kneecap you. Can you aggregate multiple gigabit connections? 2) What versions of MacOS are you using. Every MacOS release for the last few years has a different version of SMB that will or will not work well with different filer options. 3) What codecs are you using? What is your workflow? A 4k online workflow has a very different performance requirement than an offline (ProRes) workflow. 4) What is your budget? Application specific things like Avid solutions (http://www.avid.com/solutions/enterprise-media-management/media-operations) get very expensive, but are designed for these kind of situations 5) Seriously, what is your budget? 30TB of performant and somewhat resilient storage (RAID6 or RAID10) and a sensible backup solution is going to be in the $10-20k+ range at a minimum. 6) Have a back up plan as part of this. RAID isn't a backup. I used to work in post-production, and this is by no means a 'slap a Synology on the network and call it a day' situation if you have professionals who depend on this working well.
|
# ? Oct 5, 2016 17:16 |
|
mayodreams posted:Video is a special usage case of these types of devices. There are a LOT of things to consider: 2) Running Yosemite OS. We're hesitant to update it since it could cause all sorts of compatibility issues with our enterprise applications. 3) H.264 and MPEG4 4/5) $8,000 to $12,000 CAD 6) Yes, we've definitely considered this! We're looking at having a RAID setup + cloud-based backup (like CrashPlan). I realize that restoring a backup up from CrashPlan could take a very, very long time.
|
# ? Oct 5, 2016 18:59 |
|
Crashplan will be horrendous to use with many TB of video.
|
# ? Oct 5, 2016 19:29 |
|
I think the only way to really protect that amount of data is to write it out to tape and/or replicate it somewhere else. I've not really looked into how good a fit Ontap Cloud is for that sort of use case but 30TB is a poo poo ton more data than people realise.
|
# ? Oct 5, 2016 19:35 |
|
the spyder posted:I've tried every combo of logins, passwords, domain admin accounts, ect. It appears to be on the domain, as it requires a domain user account to authenticate to the shares, but I'm wondering for how long. The domain admin account was also changed, which was most likely used as the credentials for auth in OnTAP. The only note he left was to use na_admin, which 1) was discontinued in 8.1 and 2) does not work as a login in combo with any of the provided passwords. If it's joined to the domain then the Domain Admins group should be a member of the local administrators group and a domain admin account will allow you to login in that case. You would need to preface the account name with the Netbios domain name (DOMAIN\user) when logging in that way. The default administrative account is root, not na_admin. If you're logging in through the lights out management then it would be naroot, with the same password as root. The account used to join the filer to the domain is not used to process CIFS authentication. When the filer is joined to the domain a machine account is created and that machine account is used to proxy authentication requests to active directory. Your CIFS shares will continue to authenticate as long as the the time on the filer stays in sync.
|
# ? Oct 5, 2016 19:40 |
|
big money big clit posted:If it's joined to the domain then the Domain Admins group should be a member of the local administrators group and a domain admin account will allow you to login in that case. You would need to preface the account name with the Netbios domain name (DOMAIN\user) when logging in that way. The default administrative account is root, not na_admin. If you're logging in through the lights out management then it would be naroot, with the same password as root. Awesome, I can't thank you enough. As a domain admin, I am not able to use my login with the preface. The constant error is #500 connection refused, with no reason given. I'll go try the naroot account with the passwords we have.
|
# ? Oct 5, 2016 22:04 |
|
As in an HTTP error?
|
# ? Oct 5, 2016 22:15 |
|
Thanks Ants posted:30TB is a poo poo ton more data than people realise.
|
# ? Oct 6, 2016 01:50 |
|
|
# ? Apr 17, 2024 19:06 |
|
the spyder posted:Need some NetAPP advice. I just inherited a Dual Head FAS3240 with no login/password. I was able to recover the support account used and download OnCommand. It's running 8.2.4 in 7-mode. I *short-term* have console access while I'm on site. And just like the rest of my luck, it's running production critical storage. Is there ANY way I can reset the user/password without taking it offline? Oh and it literally fell out of the maintenance the day the former admin left. It was one of the very few parting bits of info he passed on. We're working to renew it/figure out what to do with it. If you're looking for cheaper maintenance to get you along for a year call up Zerowait. I am big fans of them.
|
# ? Oct 6, 2016 02:06 |