Alerting and Historical Reports for SAN data

pwyder

Hi,

I'm in the process of evaluating various SAN switch/director monitoring tools and stor2rrd is one of them.
But as of one I am missing 2 crucial elements: alerting (like alter via nagios if a port is offline for more than 2 days or if an ISl uses more than 50% bandwidth, a Power supply is defective, etc.) and history for the SAN data (possibliy with the chance of comparing load from now and 3 months before, etc.).
So my questions are: is it on your roadmap to implement these 2 features and can the tool handle more than 6000 SAN ports?

Thx 4 info
Peter

Pavel

Hi,

1. performance historical reporting is planned, hopefully next version
2. port status/event aletring is not planned. There we have no clear idea what alert. You can have many offline ports, it is nothing unusual. I am not sure if SNMP protocol provides HW events like faulty Power.
3. performance alerting is planned, definitelly next version, you will be able to define your own alert for througpuit through any port or switch.
4. you should be able to monitor 6000 ports, allocate 80GB disk space at least for database. Also start with 5 minute samples schedulled from crontab to avoid any performance issues. You can when ever change it to default 1 minute.

pwyder

Hi Pavel

Thank you for your answers.
But I do have some followup questions:
1. is apart from the historical reports/alerting for the performance planned to that also for the errors (CRC, BBZeros) ?
2. are you planning on widening the number of errory counters (eg. C3discards, EncodingIn/Out, LinkResets, etc.) ?
3. are you planning on implementing trunk-based stats and errors ?
4. I saw that you allow on the lpar2rrd part to do LDAP authentication for the logins. Is/will the same be possible for the stor2rrd part?


Thanks for the additional infos
Peter

Pavel

Hi,

1. yes, include CRC and BBzero, it is actually under development
2. no, we do not plan that as no one has ever requested that. It would not be a big issue as we have expected some enhancement here. Problem is there is many error counters, we would like to present only most important ones. we could do it if you explain us importancy of that counters.
3. I am not sure if there is directly any "trunk" based stats available. You can do "trunking" in the GUI in custom groups as a workaround.
4. LDAP will not be ported to stor2rrd. It is going to be discontinued even in lpar2rrd. Next versions of both product will have internal user management implemented, it is being deveping right now.