Tag

hard disk Archives - AWS Managed Services by Anchor

Anchor Movin’ On Out!

By | Company News | No Comments

It’s an exciting time at Anchor as another big move is in the works. We moved to the current premises at 230 Clarence about 3 years ago now. At the time we thought it was huge and had boundless plains to share, but strong business and steady recruitment has seen us grow to about fifty staff now! The exact date for the move isn’t fixed yet, but we’re prepping for the convoy to set sail sometime in late November. The destination? A few blocks east, just across the road from Hyde Park, and a substantial jump in elevation. In the meantime we’ve been doing a bit of spring cleaning, we’re getting rid of anything that we don’t need to take with us. The amount of stuff we’ve got lying around…

Read More

The difference between booting MBR and GPT with GRUB

By | Technical | 5 Comments

We recently talked about booting Linux from Really Big hard drives using GPT and a special boot partition. We thought we’d step back a bit and talk about why this is necessary, and how EFI bootloading differs from the classic BIOS boot. The case for GPT As we mentioned previously, you need to use GPT on disks larger than 2TB in size. The limit arises from the historical use of CHS addressing in the BIOS. In short, GPT lets you can create partitions larger than 2TB. MBR partitioning can’t do that. If you’re using a hardware RAID card you can carve out multiple less-than-2TB virtual disks and stick to MBR partitioning, and then glue them back together with LVM. It’s not ideal, but it works well. However, you’re out of…

Read More

Booting large GPT disks without EFI

By | Technical | No Comments

In recent times we’ve been dealing with systems with more and more diskspace. Our latest deployment has a full 30TB of usable capacity! However, this presents some interesting challenges when it comes to booting the operating system. Disks larger than 2TB necessitate the use of GPT partition tables, and booting from GPT disks involves jumping through some hoops due to the way the BIOS and bootloaders work. You can dodge this bullet if you’re using a hardware RAID card, but that’s not always an option. That’s the situation we found ourselves in the other day, with some shiny new multi-terabyte drives not attached to a hardware RAID controller. We’ll gloss over a lot of the details, but the short version is that we can’t put GRUB’s stage 1.5 in the…

Read More

Noop I/O scheduling with SSD storage

By | Technical | No Comments

Solid State Drives (SSDs) are not new in the server world, but they’re seen a somewhat limited takeup due to their high cost per gigabyte of storage. This has been changing as prices continue to drop, and we’re now at the point where SSDs represent a viable option for primary storage of high-value data that is read and written heavily. While SSDs are a drop-in replacement for traditional hard drives (HDDs) when it comes to servers, their behaviour and performance characteristics are fundamentally different. This means that the disk-access scheduling algorithms that we’ve developed over time, originally designed for rotational media, simply don’t work for SSDs. We recently observed some interesting behaviour as a result of changing disk access schedulers, which we thought we’d share with you here. The difference…

Read More

Tales of Hardware – IBM x3650

By | Technical | No Comments

All the servers Anchor buys are from Supermicro. Most people won’t have heard of them, but they’re a sizeable hardware vendor that also does some OEM gear. Supermicro certainly doesn’t carry the mindshare of other big brands like HP, Dell, et al., but we chose them because their stuff is reliable and affordable – we focus on the things that actually matter, rather than some enterprise-y idea of sticking with big brands that you trust – “noone ever got fired for buying IBM” they say. Actually, hold that thought for a moment.

Read More

A tale of two drives

By | Technical | No Comments

It’s no secret that we’d rather be working on Linux than Windows here at Anchor. It is, by and large, much more annoying to actually get anything done, but it also just breaks in opaque and unexplained ways. O Windowes, let me count the ways in which you are broken! This is one such problem we ran into yesterday. Hard drive failure is a fact of life when you run servers, by sheer virtue of that fact that you have hundreds of them. To mitigate the risk and reduce unscheduled downtime, we use Window’s built-in software RAID feature. It’s not an enterprise solution, but it gets the job done. What’s important is staying online and not losing data. Did I mention that trying to monitor a Windows box is a…

Read More