41fe77fcd45f63a661782177344c2518dc345c76 brianlee Fri Feb 14 11:42:19 2020 -0800 Small changes to Max's edits to the #hosting section on the Track Hub help page diff --git src/hg/htdocs/goldenPath/help/hgTrackHubHelp.html src/hg/htdocs/goldenPath/help/hgTrackHubHelp.html index e19dd17..9a49b82 100755 --- src/hg/htdocs/goldenPath/help/hgTrackHubHelp.html +++ src/hg/htdocs/goldenPath/help/hgTrackHubHelp.html @@ -845,89 +845,90 @@ use the "-extra=" option to specify this file when running hubCheck. For example, if you knew that a setting called "ensemblAssemblyName" was supported for use in track hubs by Ensembl, you could create a single line file that included the setting "ensemblAssemblyName". Then, when you want to check a hub that includes these extra trackDb settings, you would then specify this extra settings file on the command line:

$ hubCheck  -checkSettings -extra=http://genome.ucsc.edu/goldenPath/help/examples/hubExamples/hubCheckUnsupportedSettings/myExtraSettings.txt http://genome.ucsc.edu/goldenPath/help/examples/hubExamples/hubCheckUnsupportedSettings/hub.txt
 

(Note: The settings listed here in the "extra" file are just examples and do not represent real trackDb variables for hubs at Ensembl.)

Where to host your data?

As stated in What Are Track Hubs?, track hubs files must be located -in web-accessible locations that support byte-range requests. There are four options for these: +in web-accessible locations that support byte-range requests. Four options for hosting include:

Your Institution: Many universities provide a location for researchers to place shareable data on the web and contacting your institution's system administrators will help discover a location to store your data. For example, if you work at the NIH, there is an internal data sharing NIH network site. Sometimes institution firewall rules can change, and you may need to inform your system administrators to add browser IP addresses as exceptions, listed here. -Usually your IT department can direct you to someone +Usually, your IT department can direct you to someone who manages webspace for individual groups. This is our recommended option, as it is usually free, very fast and you can update the files yourself easily.

Webspace providers: If your institution does not provide any web hosting space for you, the most convenient solution is usually to buy a virtualized webspace server from a commercial web hosting provider. Files can be uploaded with FTP, rsync or scp and appear on a https:// domain. Avoid unlimited offers, they often do not allow binary files and are slower, rather look for a "virtual private server" (VPS). -Some exemples of providers are: A2 Hosting, BlueHost, GoDaddy, +Some examples of providers are: A2 Hosting, BlueHost, GoDaddy, HostGator, Hostinger, DreamHost, but there are many others. This is not a complete list and we do not endorse a particular one. You can search the internet for "virtual private server comparison". Offers start at around 5-10$/month for 25-50GB of storage. The advantage of VPS providers is that they bill a flat rate per month, which may be -easier to order through Universities than the per-GB billing of cloud +easier to order through Universities than the per GB transferred billing of cloud providers. For optimal performance, select a West Coast / San Francisco data center when ordering a web server, as this is closest and fastest from UCSC. There are usually no backups included, so it is good to keep local copies of your files.

Cloud providers: In general, commercial online cloud backup providers -that charge a flat rate, like Dropbox, iCloud, Google Drive, Box.com, Microsoft -OneDrive, Tencent Weiyun, Yandex.Disk, etc. do not work reliably as their business +that charge a flat rate (Dropbox, iCloud, Google Drive, Box.com, Microsoft +OneDrive, Tencent Weiyun, Yandex.Disk, etc.) do not work reliably as their business model requires rare and rate-limited data access, which is too slow or too limited for genome annotation display. However, commercial -cloud storage offers that charge per GB transferred, like Amazon S3, Microsoft -Azure Storage, Google Cloud Storage, Backblaze, Alibaba Object Store, etc. +cloud storage offers that charge per GB transferred (Amazon S3, Microsoft +Azure Storage, Google Cloud Storage, Backblaze, Alibaba Object Store, etc.) typically do work. As of 2020, they cost around 2-3 US cents per GB/month to store -the hub data and 12-18 US cents per GB transferred, when the hub is used +the hub data and 12-18 US cents per GB transferred, when the hub is used. For optimal performance, select a San Francisco / San Jose data center for the main UCSC site genome.ucsc.edu, a Frankfurt/Germany data center for genome-euro.ucsc.edu and a Tokyo data center for genome-asia.ucsc.edu. You may also want to review this discussion about issues with distributed storage servers. These services are external to UCSC and may change.

Free webspace: If you do not want to pay for web space, +and your institution does not provide a data location supporting byte-range requests, we know of at least the following sites where you can host research data and configuration files for free: Each of the providers above has a slightly different approach to hosting data for compatibility with the UCSC Genome Browser, and may have different advantages and disadvantages, such as size limitations, usage statistics, and version control integration. Additionally, as previously mentioned, any provider that supports byte-range access will work for hub hosting, and you are not limited to the above sites. Below is a summarized guide for each of the providers mentioned above.