Hi Nils, A summary what we discussed on the phone, see my replies inline
Kind Regards Detlef Nils Goroll wrote: > Hallo Detlef, > > der moderator hat mein posting wohl noch nicht freigegeben, daher > nochmal so: > > > * I suggest to also evaluate the Seconds_Behind_Master value in > check_mysql_slave > and mark the resource as DEGRADED if the replication lag is larger than a > configurable > value In contradiction what we discussed at our call I will make it an optional parameter of the MySQL resource. I will include this functionality in the MySQL agent and in the geocontrol at a later point in time. After thinking a bit longer I came to the conclusion, that this functionality should be available for MySQL agents without the geo protection as well.I filed the RFE 6739256 to track this issue. > > > * Parsing the status information can be simplified (and probably be > implemented more > robust regarding changes in the table format) by using ego (\G) to > display > results vertically > (show slave status\G) and parse the name/value pairs, e.g. in a while > read loop During the time the names were changing, so I will not gain anything. The MySQL agent relies on the system locale, and the text of show slave status \G could be affected by localisations. So I decline this one here. > > * I believe check_mysql_slave() won't work if multiple masters are > being defined > on the > slave (outside the cluster) - at least this limitation should be > documented if > it's by design. For the geo approach this is a restriction, that will be documented. > > * A general point which is not specific to your changes: Many OHAC (ex > sun > cluster) shell > scripts' robustness would probably benefit from making more use of > local variables. In the geo parts of the MySQL agent I used set -u and the problem with uninitialized variables we have been discussing is not there any more. When the complete shell scripts will be remodelled I will use set -u there as well. > > > * If I am not mistaking, the Mysql agent doesn't work with bind > addresses (bind to > (exactly one) specific address rather than IN_ADDR_ANY. I have once > developed > a workaround for this, but maybe someone wants to fix it? (I could > help out if > needed) > (this is not related to Detlef's changes either) As we discussed you were referring to an old version of the MySQL agent which was not able to run on a configured port. The bind address on the standard port was never a problem. The port problem was fixed a while ago. > > Nils > -- ***************************************************************************** Detlef Ulherr Staff Engineer Tel: (++49 6103) 752-248 Availability Engineering Fax: (++49 6103) 752-167 Sun Microsystems GmbH Amperestra?e 6 mailto:detlef.ulherr at sun.com 63225 Langen http://www.sun.de/ ***************************************************************************** Sitz der Gesellschaft: Sun Microsystems GmbH, Sonnenallee 1, D-85551 Kirchheim-Heimstetten Amtsgericht Muenchen: HRB 161028 Geschaeftsfuehrer: Thomas Schroeder, Wolfgang Engels, Dr. Roland Boemer Vorsitzender des Aufsichtsrates: Martin Haering *****************************************************************************
