Since we have this known limitation of up to 350 LVs per VG, that significantly affects vdsm performance, as documented here: https://access.redhat.com/site/solutions/441203 RHEV should give the Admin User a notification that his Storage Domain is approaching the limit and he should consider adding another storage domain to his setup. At what number of LVs should we give the notification? Maybe Storage team can help. 300? Message should include the name of the Storage Domain, of course. Request comes from GSS.
Let's have a warning on creating a snapshot (or doing LSM) after 300 volumes (confuigurable)
(In reply to Allon Mureinik from comment #3) > Let's have a warning on creating a snapshot (or doing LSM) after 300 volumes > (confuigurable) Hi LVs are not only snapshots but also templates and disks. in case the limit is 350 than it is for all kinds so warning only about snapshots isn't good enough Am I wrong? anyway it sounds like we should test only the limit - reach the limit to get the warning, doesn't matter how. we can set the limit to 5, create 5 LVs and check we get warning. please approve.
(In reply to Aharon Canan from comment #4) > (In reply to Allon Mureinik from comment #3) > > Let's have a warning on creating a snapshot (or doing LSM) after 300 volumes > > (confuigurable) > > Hi > > LVs are not only snapshots but also templates and disks. Every volume is an LV - this can be the first volume in a image (i.e., when you create a disk), or any other one (when you create a snapshot). A template itself is not an LV - its disks are. > in case the limit is 350 than it is for all kinds so warning only about > snapshots isn't good enough Yup, see above. > anyway it sounds like we should test only the limit - reach the limit to get > the warning, doesn't matter how. we can set the limit to 5, create 5 LVs and > check we get warning. We should test that every flow that adds an LV has this warning.
This RFE is verified on rc1,though the message appears only on engine's log: 2014-08-24 12:15:23,264 WARN [org.ovirt.engine.core.dal.dbbroker.auditloghandling.AuditLogDirector] (ajp--127.0.0.1-8702-7) [79950f3d] Correlation ID: null, Call Stack: null, Custom Event ID: -1, Message: The number of LVs on the domain iscsi exceeded 30, you are approaching the limit where performance may degrade. This warning could be missed quite easily,depending on the traffic of-course. Anyway,my opinion is that UI should prompt a warning message as well. what do you think,Allon?
It should appear in the alerts bar. Doesn't it?
nop,it doesn't. I have set AlertOnNumberOfLVs=30 I have almost 60 lv's now,the warning appears only on engine.log
Tal, can you take a look at this please?
I'm moving it back to ON_QA based on Allon's comment #7,please notify me if a warning on engine.log file is satisfying for a verification of this RFE...
following comment #7, reopening
Ori, I've just checked and the warning does appear in the events tab, not to mention that if you saw that AuditLogDirector line in your log, it means that it persisted an event log. Please check again, make sure you open the events tab to see all late events so a newer event won't mask out the event you are looking for
Guys, lets close this. Tal, Ori, please add screenshots of the system [not] showing the desired event - I'm guessing the message appears in place X while Ori was looking at place Y.
Created attachment 930822 [details] Events footer
Created attachment 930823 [details] Events main tab
Done, Ori please have a look
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHSA-2015-0158.html