So, I shouldn’t plan to use my 4770K for a NAS? Granted the workload it’ll see would be a fraction of what you’re doing.
💬 OT 💻 Post your computer setup
- Thread starter DanJ
- Start date
Its going to depend on what you're doing with it. I'm rocking a 4790k with no issues. It's pretty much exclusively Plex though
My Plex sees minimal usage actually.
It would be primarily a file server, as I plan to go with a new Mac next, and would be working on photos and 4K video. All hobby, nothing professional. The gigabit bottleneck concerns me.
I know I can go 10GBe on a new Mini, but everything else would need to be upgraded (switches, NIC), cables.
Taking @Thirdgen89GTA’s instructions I got a USB-C enclosure with a mid-level NVMe and bejeebus that spoiled me.
I’d like a NAS with at least that kind of transfer speeds to/from it, but I’ll have to donate a few organs to afford it.
If I had to wager a guess, I’ll probably end up with a TB3/4 enclosure and two larger NVMe drives that I’ll have to physically move around with the computer. The desktop will probably just be long term file retention and archiving.
10gb add-in cards aren't too pricey these days and then you'd have your pic of NIC chipset?
I'm looking to redo my NAS soon, I've been running ZFS on an ubuntu box built off an Asus server with a Xeon 2620v2, 64GB of ECC+Registered memory. I'm convinced it has been so problem-free BECAUSE I stuck with server-class memory. The existing server was also hosting VMs and docker containers, but last spring I built an all-flash vSAN cluster with 3 epyc 7302p's with 128gb of ram each so I've been transferring compute to kubernetes running on top of the vSAN cluster. My NAS will just need to handle storage duties but I want it to be FAST and HUGE. Still trying to figure out what I'm going to do there.
I'm looking to redo my NAS soon, I've been running ZFS on an ubuntu box built off an Asus server with a Xeon 2620v2, 64GB of ECC+Registered memory. I'm convinced it has been so problem-free BECAUSE I stuck with server-class memory. The existing server was also hosting VMs and docker containers, but last spring I built an all-flash vSAN cluster with 3 epyc 7302p's with 128gb of ram each so I've been transferring compute to kubernetes running on top of the vSAN cluster. My NAS will just need to handle storage duties but I want it to be FAST and HUGE. Still trying to figure out what I'm going to do there.
Today we had to restore multiple drive volumes, about 12TB of data all told.
So, how do you quickly sequentially copy all of the different folders to different destinations easily?
The young Padawan was doing to drag the contents folder to the new drive, wait for it to finish, then do the next drive letter. And there were about 9 folders. Each folder represented a volume, d:, e:, f:...etc.
No my Young Padawan. That is the HARD way.
Here is the SMART way.
Batch may be old, but the FOR command is still infinitely useful for repetitive tasks. Hit enter, and walked away for about 50 hours so the copy can complete.
So, how do you quickly sequentially copy all of the different folders to different destinations easily?
The young Padawan was doing to drag the contents folder to the new drive, wait for it to finish, then do the next drive letter. And there were about 9 folders. Each folder represented a volume, d:, e:, f:...etc.
No my Young Padawan. That is the HARD way.
Here is the SMART way.
Code:
for %g in ( j k l n q r s t u ) do ( robocopy x:\%g %g:\ /mir /z /mt:16 /j )Batch may be old, but the FOR command is still infinitely useful for repetitive tasks. Hit enter, and walked away for about 50 hours so the copy can complete.
So, was looking at the TrueNAS server today, specifically used capacity.
I took a peek at the "Masters" folders I keep on each Media volume. The "Masters" folder is where I dump raw files before they get named, sorted, and eventually filed into the correct Media folder. Its also where I put the special features for TV shows and movies before I run them through final transcoding. As they are special features, I don't see a need to keep them around in full quality.
I might have to spend some time going through the Masters folder. Combined, both folders are sucking up 880GB worth of space. Whoops.
I took a peek at the "Masters" folders I keep on each Media volume. The "Masters" folder is where I dump raw files before they get named, sorted, and eventually filed into the correct Media folder. Its also where I put the special features for TV shows and movies before I run them through final transcoding. As they are special features, I don't see a need to keep them around in full quality.
I might have to spend some time going through the Masters folder. Combined, both folders are sucking up 880GB worth of space. Whoops.
Code:
Asgard% du -sh /mnt/Frigga/Media2/Masters /mnt/Odin/Media/Masters
79G /mnt/Frigga/Media2/Masters
803G /mnt/Odin/Media/Masters
Thats not socially appropriate. Rename your folders. 
My i5 6500 is still plenty right now for my unRaid server with plex. With my wireless being spotty at best lately, I am looking into either a mesh set up, or get used enterprise shtuff. The TP Link Deco X20 is looking like it should suit needs for now. Or should I splurge and get a nice multiport switch and some wifi 6 access points.
My i5 6500 is still plenty right now for my unRaid server with plex. With my wireless being spotty at best lately, I am looking into either a mesh set up, or get used enterprise shtuff. The TP Link Deco X20 is looking like it should suit needs for now. Or should I splurge and get a nice multiport switch and some wifi 6 access points.
So here's a fun thing.
At work, I need to move a whole bunch of files from multiple servers, to a single share. Not so hard if you have a list of files in a UNC format. Easy peasy.
How about when the idiot gives you the files in the TREE ASCII format like this:
cmatskas.com
So, I have a text file that uses ASCII characters, and I need to convert that back to UNC network paths.
Not so bad if they are all on one server, and not too many files.
How about 450,000 lines?
Anyone know of code functions that can do this without having to write one completely from scratch? I'm not having fun with this one.
At work, I need to move a whole bunch of files from multiple servers, to a single share. Not so hard if you have a list of files in a UNC format. Easy peasy.
How about when the idiot gives you the files in the TREE ASCII format like this:
Generate ASCII folder structures on Windows with Tree
I was reading through Adron Hall's post yesterday about generating ASCII trees on Linux and Mac OSx. This is a handy feature if you want to include a project/folder structure to your documentation and don't wish to copy paste screenshots. What is an ASCII tree you may ask? There
Code:
+---.atom
| \---compile-cache
| \---836a81e7afc5e1080850f10b0790220f3d81503a
| \---js
+---.p4merge
+---Contacts
+---Desktop
+---Documents
| +---Custom Office Templates
| +---Fiddler2
| | +---Captures
| | | +---Requests
| | | \---Responses
| | \---Scripts
+---Favorites
| +---Links
+---Links
+---Music
+---Pictures
+---Saved Games
+---Searches
+---Source
| \---Repos
\---VideosSo, I have a text file that uses ASCII characters, and I need to convert that back to UNC network paths.
Not so bad if they are all on one server, and not too many files.
How about 450,000 lines?
Anyone know of code functions that can do this without having to write one completely from scratch? I'm not having fun with this one.
So, I basically went dark, still busy as fuck with an project at work.
But, I did what was thought was impossible after nearly 3 days of trying various methods, now, the task of actually implementing it. Dry run was not promising. Speed is a problem. Need to rewrite portions of the script to reduce memory/disk IO footprint.
Doesn't help that the error log was over 30MB in size.
I had originally been holding the errors in a custom Powershell object so I could manipulate the data later as needed and act against it to handle errors intelligently. That method is definitely not going to work. I need to cut the Powershell array down to bare minimum's to keep its memory footprint down, and write everything else out to a transaction log in CSV format.
The good news, I already know how to do all of that already. I'm annoyed that I have to though. Transaction logs are a pain if I need the script to do anything with it post process to automatically handle certain things. And I didn't want to get too complex with the functions because this project is essentially a one and done. Everything I do here, will get used for the project, and then never used again.
On the FUN side of things.
The SN850 2TB SSD's are dropped in price MAJORLY.
$270 on Amazon! Gonna look at the reviews and see if anything has changed. I seem to remember the KC3000 and Seagate Firecuda 530 getting some good reviews.
WD_BLACK 2TB SN850 NVMe Internal Gaming SSD Solid State Drive - Gen4 PCIe, M.2 2280, 3D NAND, Up to 7,000 MB/s - WDS200T1X0E
by Amazon.com
Learn more: https://www.amazon.com/dp/B08KFRFL8F/?tag=tcg21-20
But, I did what was thought was impossible after nearly 3 days of trying various methods, now, the task of actually implementing it. Dry run was not promising. Speed is a problem. Need to rewrite portions of the script to reduce memory/disk IO footprint.
Doesn't help that the error log was over 30MB in size.
I had originally been holding the errors in a custom Powershell object so I could manipulate the data later as needed and act against it to handle errors intelligently. That method is definitely not going to work. I need to cut the Powershell array down to bare minimum's to keep its memory footprint down, and write everything else out to a transaction log in CSV format.
The good news, I already know how to do all of that already. I'm annoyed that I have to though. Transaction logs are a pain if I need the script to do anything with it post process to automatically handle certain things. And I didn't want to get too complex with the functions because this project is essentially a one and done. Everything I do here, will get used for the project, and then never used again.
On the FUN side of things.
The SN850 2TB SSD's are dropped in price MAJORLY.
$270 on Amazon! Gonna look at the reviews and see if anything has changed. I seem to remember the KC3000 and Seagate Firecuda 530 getting some good reviews.
WD_BLACK 2TB SN850 NVMe Internal Gaming SSD Solid State Drive - Gen4 PCIe, M.2 2280, 3D NAND, Up to 7,000 MB/s - WDS200T1X0E
by Amazon.com
Learn more: https://www.amazon.com/dp/B08KFRFL8F/?tag=tcg21-20
Replaced my Asus router that either didnt like being in storage, or just wanted to not do its job to the potential it had anymore. Went with TP Link Deco X60 mesh setup. The main Deco by itself was doing more work than the Asus was, but if I paid for a 2 pack, Im gonna use both god damn it!!
400mbps in my garage? SURE!!
Ring Doorbell still looks like crap and its self health diagnostics still show weak wifi. With the second deco less than 10 feet on the other side of the wall. I gave this thing every shot at succeeding, but I think its time to upgrade. Id go with Wyze, but no SD card option and just like Ring, you need to buy their subscription to see previous recordings.
400mbps in my garage? SURE!!
Ring Doorbell still looks like crap and its self health diagnostics still show weak wifi. With the second deco less than 10 feet on the other side of the wall. I gave this thing every shot at succeeding, but I think its time to upgrade. Id go with Wyze, but no SD card option and just like Ring, you need to buy their subscription to see previous recordings.
I saw this thread initially and thought it would be just gaming computers, so I didn't even look. I have what used to be a decent machine for gaming but I really don't get much time or joy from it, so when I do play anything it's borderlands on PS4.
However I do have a baby homelab. Lenovo tiny thinkcentre, i3 8th gen running Plex. Internal m.2 OS drive, 1 tb storage for steam and a home network file share. I picked up a 12tb wd drive last year that Plex is pointed to. That combined about 4 external hard drives of media into a central location with about 8tb to spare, I guess I need to look into downloading more Linux ISOs.
I also have a couple pi's, one of which is my pihole and NTP for my R420 when it's running. Im not rich enough to have it running for no reason all the time. I've been using it lately to try to learn hyper v, because all of my experience is with VMware. The second pi is used for Linux learning, which is my biggest weakness.
I've been toying with the idea of putting proxmox on my thinkcentre to host a Linux box and a Windows box at once, but I'm not sure about passthrough for the external drive. We may be about to move so that may be an option when I'm setting everything up at the new place.
However I do have a baby homelab. Lenovo tiny thinkcentre, i3 8th gen running Plex. Internal m.2 OS drive, 1 tb storage for steam and a home network file share. I picked up a 12tb wd drive last year that Plex is pointed to. That combined about 4 external hard drives of media into a central location with about 8tb to spare, I guess I need to look into downloading more Linux ISOs.
I also have a couple pi's, one of which is my pihole and NTP for my R420 when it's running. Im not rich enough to have it running for no reason all the time. I've been using it lately to try to learn hyper v, because all of my experience is with VMware. The second pi is used for Linux learning, which is my biggest weakness.
I've been toying with the idea of putting proxmox on my thinkcentre to host a Linux box and a Windows box at once, but I'm not sure about passthrough for the external drive. We may be about to move so that may be an option when I'm setting everything up at the new place.
Seeking some help from the IT guru's here. My friend is going through an IT course and needs to write up a little essay on an service attack. I have no experience or dealings with them, so if anyone wants to weigh in with a example or experience with one, it would be much appreciated.
These are the questions associated with the required essay:
__________________________________________________________________________
Case Project 1-1 Personal Attack Experiences
What type of computer attack have you (or a friend or another student) experienced? When did it happen? What type of computer or device was involved? What type of damage did it inflict? What had to be done to clean up following the attack? How was the computer fixed after the attack? What could have prevented it? List the reason or reasons you think that the attack was successful. Write a one-page paper about these experiences.
Use Microsoft Word to submit Case Project 1-1. Make sure to answer all the questions.
Writing is essential for IT professionals. Both writing and technical continent will be graded.
These are the questions associated with the required essay:
__________________________________________________________________________
Case Project 1-1 Personal Attack Experiences
What type of computer attack have you (or a friend or another student) experienced? When did it happen? What type of computer or device was involved? What type of damage did it inflict? What had to be done to clean up following the attack? How was the computer fixed after the attack? What could have prevented it? List the reason or reasons you think that the attack was successful. Write a one-page paper about these experiences.
Use Microsoft Word to submit Case Project 1-1. Make sure to answer all the questions.
Writing is essential for IT professionals. Both writing and technical continent will be graded.
Go for it. She's basically at the end of the first chapter/level/course, whatever, so it doesn't have too be too detailed. I wouldn't think the teacher can expect a real granular thesis considering they just started the course.
What type of computer attack have you (or a friend or another student) experienced? Email server breach. Look up proxy logon and proxy shell. After the initial proxy logon, servers were patched. Proxy shell was a separate attach, and overlooked by our security as a modified version of proxy logon. IDS and IPS show attempts and blocking scripts. However, a single user opened the attachment and enabled access though a malware infected excel spreadsheet. You can also look up squirrel waffle, and see the malware attachment.
When did it happen? August was the first notice of draft emails in users inboxes. Early November is when the attackers referenced stolen emails to customers and other users to send malware.
What type of computer or device was involved? Virtual machine hosted on vmware environment.
What type of damage did it inflict? Attackers were able to gain access to emails that included customer information, credit card, and passwords that were sent unencrypted.
What had to be done to clean up following the attack? Company's insurance provided a cyber security firm to analyze the compromised server and provide data that the attackers had. This enabled the company to provide customers with compromised PII
How was the computer fixed after the attack? Wiped. Gone. Replaced and hardened. Company decided against on premises email server hosting and we are currently migrating to office 365 hosted application.
What could have prevented it? Manpower. It was an oversight of one person. Having a team of qualified individuals instead of one overworked person identify and remediate problems could have prevented the attack. Following the attack, three new positions were opened, a security, a systems engineer, a part time help desk technician were opened. The company learned that information technology security should not be an option but a priority.
List the reason or reasons you think that the attack was successful. Opportunity and availability. Attackers will use any vector for attack. When patches for vulnerabilities are released, attackers also gain the knowledge of exploits. They can easily probe for available avenues to attack and any openings are then further investigated, until the find some poor sap with bad patch management.
When did it happen? August was the first notice of draft emails in users inboxes. Early November is when the attackers referenced stolen emails to customers and other users to send malware.
What type of computer or device was involved? Virtual machine hosted on vmware environment.
What type of damage did it inflict? Attackers were able to gain access to emails that included customer information, credit card, and passwords that were sent unencrypted.
What had to be done to clean up following the attack? Company's insurance provided a cyber security firm to analyze the compromised server and provide data that the attackers had. This enabled the company to provide customers with compromised PII
How was the computer fixed after the attack? Wiped. Gone. Replaced and hardened. Company decided against on premises email server hosting and we are currently migrating to office 365 hosted application.
What could have prevented it? Manpower. It was an oversight of one person. Having a team of qualified individuals instead of one overworked person identify and remediate problems could have prevented the attack. Following the attack, three new positions were opened, a security, a systems engineer, a part time help desk technician were opened. The company learned that information technology security should not be an option but a priority.
List the reason or reasons you think that the attack was successful. Opportunity and availability. Attackers will use any vector for attack. When patches for vulnerabilities are released, attackers also gain the knowledge of exploits. They can easily probe for available avenues to attack and any openings are then further investigated, until the find some poor sap with bad patch management.
So, I've been looking at docks to expand my Mini's ports, I love the machine, but man, a few more TB3 ports, or even more USB-A ports would have been nice. So I've been using a Dell USB-C dock from work that's for my laptop at home.
Ran across this though. Slightly disappointed at only 3 USB-A ports, and 1 USB-C. But, the two M2 bays would be nice to have an external drive hooked up to the Mini without the extra cable mess.
Ran across this though. Slightly disappointed at only 3 USB-A ports, and 1 USB-C. But, the two M2 bays would be nice to have an external drive hooked up to the Mini without the extra cable mess.
ORICO Thunderbolt 3 M.2 Dual-bay Docking Station
ORICO Thunderbolt 3 M.2 Dual-bay Docking Station
www.orico.cc
She can say whatever she wants. Maybe she can just write about how her dear aunt got phished and the hacker got into her e-mail or something. Or can say that she accidentally downloaded some adware and had to get it cleaned up. it would be hard to not get a one page paper out of that.
Anyone hear about this?
www.macrumors.com
Western Digital says that it has lost 6.5 exabytes worth of BiCS 3D NAND flash storage due to use of contaminated materials at two of its plants in Japan in January. The facilities are jointly run by Western Digital and its partner Kioxia, which was previously part of Toshiba. What caused the contamination and when production is expected to resume has yet to be confirmed.
Western Digital and Kioxia are among the world's leading SSD suppliers, with their joint production amounting to around 30 percent of the entire NAND flash market. TrendForce's market research indicates that the price of NAND, which is the main component of SSDs, could spike by up to 10 percent starting in the second quarter of 2022.
Western Digital and Kioxia primarily offer SSD and eMMC storage drives for PCs, but iFixit teardowns have shown that Apple uses Kioxia NAND chips in many of its products, including the iPhone 13 lineup, the iPad Pro, and the 14- and 16-inch MacBook Pro. Whether any existing Apple products manufactured with Kioxia storage in January will need to be recalled is currently unclear.
Many of Apple's Mac SSDs have been supplied by Samsung in recent years, but the company has increasingly turned to custom storage solutions, culminating in the proprietary SSD components on Apple's ‌M1‌ Macs being soldered in place.
If Apple devices are not directly impacted by contaminated Kioxia hardware, it is still possible that a tightening of supply across the industry could drive up Apple's costs, but whether the company will absorb these price rises or pass them on to consumers is yet to be seen.
Apple SSD Supplier Hit With Contamination Disaster, Flash Storage Prices Expected to Spike
SSD prices are likely to increase globally due to contamination issues at Western Digital and Kioxia's NAND production facilities, and Apple...
www.macrumors.com
Western Digital says that it has lost 6.5 exabytes worth of BiCS 3D NAND flash storage due to use of contaminated materials at two of its plants in Japan in January. The facilities are jointly run by Western Digital and its partner Kioxia, which was previously part of Toshiba. What caused the contamination and when production is expected to resume has yet to be confirmed.
Western Digital and Kioxia are among the world's leading SSD suppliers, with their joint production amounting to around 30 percent of the entire NAND flash market. TrendForce's market research indicates that the price of NAND, which is the main component of SSDs, could spike by up to 10 percent starting in the second quarter of 2022.
Western Digital and Kioxia primarily offer SSD and eMMC storage drives for PCs, but iFixit teardowns have shown that Apple uses Kioxia NAND chips in many of its products, including the iPhone 13 lineup, the iPad Pro, and the 14- and 16-inch MacBook Pro. Whether any existing Apple products manufactured with Kioxia storage in January will need to be recalled is currently unclear.
Many of Apple's Mac SSDs have been supplied by Samsung in recent years, but the company has increasingly turned to custom storage solutions, culminating in the proprietary SSD components on Apple's ‌M1‌ Macs being soldered in place.
If Apple devices are not directly impacted by contaminated Kioxia hardware, it is still possible that a tightening of supply across the industry could drive up Apple's costs, but whether the company will absorb these price rises or pass them on to consumers is yet to be seen.
Holy shit. Is exabyte one or two above petabyte? Cause either way that’s a LOT of wasted SSDs.
From a different article:
Western Digital and Kioxia revealed this week that contaminated materials used in flash-memory chips ruined at least 6.5 exabytes, or more than 6.5 million terabytes, of storage.
My maths sucks, but that seems less than what
Thirdgen89GTA has on his PMS.
Western Digital and Kioxia revealed this week that contaminated materials used in flash-memory chips ruined at least 6.5 exabytes, or more than 6.5 million terabytes, of storage.
My maths sucks, but that seems less than what
Thirdgen89GTA has on his PMS.
Even so, he uses, along with most of us, spiny rust drives. Not 3D NAND flash drives. That’s probably what makes this worse.
Okay, so you want to know the basics of DoS mitigation the first case is a one off, interesting example. I call this "the sword of SEO." It would take 2X4 core server with 24 gigs RAM and bring it to its knees in a few minutes. How was it done?
The first thing was to get traffic to go to a specific web page controlled by the attacker. This was set up on a free hosting web site.
The second thing was to direct traffic there. This guy had some sort of SEO skills to do that.
The third thing was to put 30 or so iframes on the page to hit the person's source server. This was targetting a non-static page (e.g. the front page of their forum) so that it would generate as many non cached DB hits as possible
By flooding the system with non-cached DB queries it just beat it to death, as said very quickly. After the discovery of this we set up a rule in the web server that blocked the page that had the iframes on it (they showed as referrals.) Problem handily solved.
In terms of the "common" DoS attack you would have a source or sources send UDP fragments to a host with a malformed source. This makes it virtually impossible to know what the real source is and unless you block UDP upstream almost impossible to block. It works even better if you can make a query to someone else's computer and get them to send out more data to your victim than the original query had in it (amplification.) Virtually any and all hosting providers who have such an attack impact their network WILL NULL ROUTE THE TARGET. Notice I said the target, not the source. This is super common in the industry and I've seen anything from game servers to school supply companies get hit with this by the way.
Thread Info
-
Created -
3KReplies
-
84KViews
-
Participants list
