Extracting a web site from a client's host

This is a multi-step process to extract the database and site files from a client’s hosting account. We have to be extremely careful as these are production websites and we cannot make any modifications to the site on their server.

Initial Evaluation of their site. We do not know what the status is of their website, there are times where the database or the site files may be huge, beyond a reasonable size. For example, the directory for a website could be 30GB or more. This indicates a problem so we need to address this before we can extract the site.

Login to the cpanel using the info that we have provided.

Check for a way to get SSH access, this will vary from host to host, please use Google to search for how to get this information for the particular host.

If you cannot find a way to get SSH access, then you need to update Wrike with this info and ping the manager your are working with, we will then need to discuss with the client the next option.

If you can get SSH access, then login via SSH and run the following commands, then report update the results in Wrike:

du -hs /*| sort -h

If any folders seem unusually large, ‘cd’ into the and then run this again for that folder.

du -hs /*| sort -h

Keep doing this you can figure out which folder or folders actually are the root of the problem, then update Wrike with this info.

2) Do not proceed on this step until the manager has approved the results of step 1 in Wrike. The reason for this is that the site may be too large to where there are major issues that need to be addressed first. In this case, we may need to ask the client to fix this first before moving to this step.

If step 1 was approved with no issues, then proceed to do a regular tar compression of the site folder directory If step 1 was determined to be too large, do a modified tar compression by excluding certain folders or files that were determined to be ok to leave out, example:

tar -zcvf /all.tgz / --exclude=

3) Once the tar file is ready, scp it to our server:

Posted on Dec 04, 2017