MySQL table keeps ballooning...has anyone seen this?

prodservices
New Contributor III

We've got a table in our MySQL database that had ballooned to a rather large size that we discovered prior to our last upgrade. The database itself was reasonable, but the "downloadable_file_chunk_data" table was enormous. The solution was to run the following in MySQL:

use jamfsoftware
then
truncate downloadable_file_chunk_data
(reboot tomcat, then)
/usr/local/mysql/bin/mysqlcheck -u root --optimize jamfsoftware

This resolved our issue and the backup became manageable again, upgrade to most current version went fine. Checking our backups though we saw this:

2.7G -rw-rw----. 1 mysql mysql 2.7G Nov 18 13:10 applications.MYD
11G -rw-rw----. 1 mysql mysql 11G Nov 17 14:08 downloadable_file_chunk_data.MYD
6.2G -rw-rw----. 1 mysql mysql 6.2G Nov 18 13:10 downloadable_file_chunk_data.TMD

So whatever mitigated this is still there. Although it's easily fixed I would love if someone who's been using Casper for a while and has a better understanding of MySQL could shed light on what's causing this in the first place. I prefer to be proactive not reactive. I found some info on the nation suggesting packet size could play into it but not being a MySQL guru I can't say.

If it's relevant we're hosting our JSS/JDS on RHEL and it's currently an overpowered physical machine for the size of our environment (64 gb RAM / 5 tb storage). Much thanks to any who have seen this and can point me in the right direction.

4 REPLIES 4

Jookyseacap
New Contributor III

That table is used for uploading files through tomcat to the JDS servers. This post by @ShaunM9483 does a great job of explaining the process. https://jamfnation.jamfsoftware.com/discussion.html?id=10730

were_wulff
Valued Contributor II

@prodservices

That table ballooning and shrinking in size normal and expected behavior when a JDS is in the environment.

Once the transfer between JDSes or the upload to the JDS from Admin is complete, the table should clear itself out.
If it does not, it means something got stuck or failed and isn't restarting and clearing out properly; in those cases, support will usually walk you through truncating that table and, if a replication is supposed to happen, the JDS will try again.

As long as a JDS or JDSes are in play in your environment, you will see this table balloon and shrink as files are replicated between the JDSes.

If you are using JDSes and are consistently seeing this table balloon in size and not shrink back down, please get in touch with your Technical Account Manager so they can help dig into it a bit deeper and find out why it's not clearing out as expected.

Thanks!
Amanda Wulff
JAMF Software Support

ryanstayloradob
New Contributor III

Ah, that makes sense now. Back when we used JDS's, the table ballooned and didn't shrink. We had other issued wit the JDS's and moved back to FDS and have not had any issues since.

lkrasno
Contributor II

The optimize also occurs when a backup is ran, do you have them scheduled?