Tag Info

Hot answers tagged

9

Cost, capacity and reliability are factors for why SSD adoption hasn't occurred at all levels. SSDs cost more than SAS disks for a given capacity. But in general, servers don't actually come with a particular type of disk. Storage is something that is configured afterwards. Some background information: Are SSD drives as reliable as mechanical drives ...


5

In many cases, enterprise storage frames actually employ three different kinds of disks- SSD, SAS and SATA (or LSAS drives which are SATA with a SAS controller), with the goal being to optimize the data being stored with the I/O it actually needs. In other words, data accessed frequently ends up living in the highest speed areas (flash) and data accessed ...


4

Many servers still use spinning disks with a SAS interface, but most (all?) servers you buy these days should have options for SSDs as well. The price comparison you're making is a little tricky - an enterprise SAS disk is definitely more expensive than most consumer SSDs, but an enterprise SSD is a whole lot more expensive than a consumer one - they're ...


3

21GB out of ~6TB available => <1% Freespace. ZFS recommends 20% freespace for RAIDZ, and at least 10% is mostly mandatory for any reasonable performance. You need to free up some space or expand the size of the array. Side nodes: SATA drives need to be scrubbed weekly if you expect to detect array failures before you get into likely data-loss ...


2

Your partitions are misaligned and it might be difficult to asses how much you're actually losing in performance because that would depend on the type of I/O workload. It might be negligible if your I/O workload is light, compared to the performance of your disks. However, since this is a NFS server, I'm assuming it's not negligible and should be addressed. ...


2

Yes, this is documented on Wikipedia. Some advanced disk-drive commands, such as Native Command Queuing (which may increase performance), Secure Erase (which allows all data on the drive to be securely erased) and S.M.A.R.T. (accessing indicators of drive reliability) exist as extensions to low-level hard-drive command sets such as SCSI, SATA, or PATA. ...


1

If your application is loading lots of content, the problem may arise from low I/O rate. If you're using a small disk (I believe the smallest default, assuming you're not using a SSD, is 10GB), then you're severely limiting your IOPS and throughput since those are measured on a per-GB basis. See "persistent disk performance" in the GCE disk docs. I'd ...


1

Some monitoring and graphing tool would be useful to pinpoint the location of bottleneck. We use http://munin-monitoring.org/ Munin grabs stats from most of known services plus system metrics and it works almost out-of-the-box on modern distros.


1

I think pm.max_children = 100 is way too small for 60k concurrent visitors. Even if a visitor clicks a link only once in a minute on average then you'd need to generate 1000 page views in a second (60000/60). You have 12GB of unused memory — set for example pm.max_children = 500. Maybe even higher - depending how much RAM is needed to generate an average ...


1

Some years ago I started writing a storage system similar to ceph. Then I discovered ceph and what they had worked better so I dumped my development. During the development process I asked a similar question to yours but on SA I did a lot of calculation on handling lots of small files and found that naming files (assuming they can be anything) by uuid and ...


1

You don't provide the Windows version that you will use. I really recommend using 2012 R2 to get all new feature from NTFS, like hot repair. Your 3 nightmares will be: -Fragmentation -Time taken to do a chkdsk. The time of it is based on number of files, not size. -backup time If you are at least on Windows 2012, you should look at ReFS. This new file ...


1

Your InnoDB database buffer pool is about 6,6 gigabytes. When inserting data to the database, InnoDB engine first adds the data to the buffer pool, and then writes it to the disk. So, what happened here is that when the buffer pool got full, InnoDB started to write its contents to the disk. Therefore the insert performance dropped. Now, when you stopped ...


1

maybe you are running into issues with MTU on the connection. take a look at the following two links, they may contains some usefull info. Configure the MTU with Fragment and MSS Optimizing performance on gigabit networks


1

I have solved it. Or rather, I know how to make it faster, but do not know why it is slow. It seems that AES cipher interferes somehow with TCP on this particular server and results in the low performance. If I change the cipher to, say, 3des (which is slower as reported by "openssl speed"), I get 150mbps. After some trials, I now use this: cipher ...


1

After digging through the documentation (based on the other answers here), this is the process I ended up using: Start xperf tracing xperf -on base+interrupt Note, you will need to close Process Monitor or any other app which uses ETW or you will get the following error: xperf: error: NT Kernel Logger: Cannot create a file when that file already exists. ...



Only top voted, non community-wiki answers of a minimum length are eligible