Feature request: more granular settings of which errors generate e-mail notifications

[budy]

Hi all,

I'd like to request a feature which lets operators decide which "errors" should generate an e-mail notification.

IMHO, it is not possible to get a message if a snapshot has been assigned and thus the replication got interrupted - one would expect that if volume replication is set up and there're automatic snapshot schedules configured for that volume.

Right now, I am receiving 8+ messages each day and people tend to only skim those if they get them on a regular basis.

So, either silence those events that are issued by any schedule or at least provide a switch for the operator do do so.

Or am I the only one, who has to struggle with more e-mails than he can handle?

Cheers,
budy

[Ke-S]

IMHO, it is not possible to get a message if a snapshot has been assigned and thus the replication got interrupted - one would expect that if volume replication is set up and there're automatic snapshot schedules configured for that volume.

Can you describe this part of your post a little bit more. We don't want to add any new task when we don't fully understand the subject in the point of user. We need to know what exactly you want to get in your mail in what situation. What you mean automatic snapshot schedule in case of volume replication ?

[budy]

Hi Kent,

sorry, now that I re-read that - that was a complete mistake on my part. I should not read "possible" but "necessary".

I was referring to the loads of automatic notifications that are gererated by the scheduled snapshots on a replicated volume. So, a cycle of these "error" messages is:

- (Master) Replication:Volume Replication: Device lvasterix0101: Connection (from Connected to WFBitMapS)
- (Slave) Replication:Volume Replication: Device lvasterix0101: Connection (from Connected to WFBitMapT)
- (Slave) ... multiple errors depending on how many snapshots one runs on these replicated volumes

This is of course, what you'd expect when setting up volume replication and snapshots on the replicated volume, so these are no errors, since this is expected behaviour.

Does this sound reasonable? For me it would make a lot of sense to be able to turn those notifications off, or at least have them flagged as such, since those are not errors per se.

Cheers,
budy

[budy]

To put that into perspective: I have got 32 warnings over the last 3 days, each of which is of course an expected notification, since they all are related to the snapshot tasks, that are assigned on the slave volume.

The problem is, and I bet that you'll be with me on that: if one gets repeated warnings or errors that aren't really errors, people tend to just skim over them and are risking to miss the one important error, that they really should have paid attention to.

If I am getting 8 e-mails a day from my DSS hosts, I will start to think that everything is okay, solely based on the number of messages I get. We all know that this shouldn't happen, but if a system runs reliable for a period of time, I guess nobody can help him (or her) making that shortcut - especially if you have to take care of dozens of such systems, which all fire messages at your inbox.

Cheers,
budy

[Ke-S]

I have forwarded your suggestion to our developers. I hope it will implemented soon.

I don't know if this is possible in your mail software but maybe you can configure some filters for those mails so it will be marked as read or removed and you will be noted only about more important mails for you. It will be a good solution until we will have it implemented in our software.