Is there any contraindication in marking the radacct.AcctSessionId column UNIQUE?
From what I've read by researching in old threads, and looking at the current source [1], it looks like there is no one.
I'm still using old FR 1.1.7. I know this is a sympthom of something else, and I already investigated and found the solution, but while I wait I get this activated I need to mitigate the problem. thanks [1] https://github.com/FreeRADIUS/freeradius-server/blob/master/raddb/sql/mysql/... -- Lorenzo Milesi - lorenzo.milesi@yetopen.it GPG/PGP Key-Id: 0xE704E230 - http://keyserver.linux.it
On 15/11/12 11:51, Lorenzo Milesi wrote:
Is there any contraindication in marking the radacct.AcctSessionId column UNIQUE?
Yes. It's not guaranteed to be unique. At most, you might want to make the combination (nasipaddress, acctsessionid) unique, but only if you KNOW your NAS won't ever generate duplicate acctsessionid values.
From what I've read by researching in old threads, and looking at the current source [1], it looks like there is no one.
I'm still using old FR 1.1.7.
Upgrade.
Yes. It's not guaranteed to be unique.
So why in the latest releases it's set to be?
Upgrade.
I know :/ -- Lorenzo Milesi - lorenzo.milesi@yetopen.it GPG/PGP Key-Id: 0xE704E230 - http://keyserver.linux.it
On 15 Nov 2012, at 13:44, Lorenzo Milesi <maxxer@ufficyo.com> wrote:
Yes. It's not guaranteed to be unique.
So why in the latest releases it's set to be?
Uh it's not. acctuniqueid != acctsessionid https://github.com/arr2036/freeradius-server/blob/master/raddb/sql/mysql/sch... acctuniqueid is set to be unique, as that's *meant* to be a unique value generated by the server which is a hash of Acct-Session-ID and other attributes. -Arran
Uh it's not. acctuniqueid != acctsessionid
Yeah, my mistake :( thanks -- Lorenzo Milesi - lorenzo.milesi@yetopen.it GPG/PGP Key-Id: 0xE704E230 - http://keyserver.linux.it
participants (3)
-
Arran Cudbard-Bell -
Lorenzo Milesi -
Phil Mayers