iPhone 2.0 Battery Life and ActiveSync (Bug 29903)

[jjzhuang]

Originally Posted by Jason Hung

I agree with Matthew. J.J., we're happy Zimbra customers that are just trying to help improve the product. We think there are issues with the beta, and we're just reporting bugs as we see them. I've noticed that Bugzilla has made some changes to the bug I filed.

I haven't got time to run the fresh Zimbra installation experiment yet, but I've tested the iPhone on our decommissioned Exchange 2003 server and can say that ActiveSync works as intended with Exchange 2003.

Absolutely. We really appreciate user input. You guys are spending your own time trying to improve our product. We'll definitely look into the issue. Thanks!

[jjzhuang]

Originally Posted by mporter

FWIW, we are suffering from a Zimbra ActiveSync issue for mail. If we activate mail ActiveSync, load on the Zimbra server spikes through the roof.

Not sure if you are hitting Bug 29899 – iPhone 2.0 + Zimbra Mobile Sync can cause high CPU usage for large mailboxes

[JHill]

Originally Posted by jjzhuang

Not sure if you are hitting Bug 29899 – iPhone 2.0 + Zimbra Mobile Sync can cause high CPU usage for large mailboxes

This one's killing us... Really, really looking forward to 5.0.9.

[dNb]

Originally Posted by jjzhuang

We'll definitely look into the issue. Thanks!

Hi J.J.-

I've been paying close attention to this thread and the accompanying bug report as the iPhone 2.x adoption rate starts to pick up steam for us. I'll be doing my own testing to see if I see this bug.

I did have one question that hasn't been clearly answered from either your posts here or in bugzilla: are you folks seeing the reported battery drain back at Zimbra HQ? I know several of the engineers there besides the Mobile team seem to carry iPhones. Are they seeing this problem? Does your personal phone drain faster?

The bug reports make it seem like anyone who tries to sync Contacts/Calendars will see a big battery drain. I'm just trying to understand how prevalent the issue is (does it happen to some people with certain data, everyone, you?) before I tell people to use/avoid it for the time being.

Thanks for all of your hard work. The Mobile part of Zimbra is definitely making everyone here very happy (at least in anticipation).

-- dNb

P.S. I have no trouble believing that there is a bug in the iPhone 2.x firmware that Zimbra is tickling but Exchange does not (or it works around). It would be interesting to compare packet traces from Jason from his two test environments.

[motuman]

Edge iPhone, OS 2.0 using hosted Zimbra.

I am have the same battery troubles. It is unexpectable to disable Contact and Calendar syncing to possibly solve this problem.

The static on my speakers is constantly going off like when my non 3g Windows Mobile phone used to do its query for Exchange email.

This leads me to believe like all the other posts, activesync is in a continual state of communication with the Zimbra server causing massive battery drain.

This is a deal breaker for me. I will move to hosted Exchange account if this cannot be solved.

[dNb]

BTW, I should mention that I run the group here who controls the horizontal and the vertical in our environment (the network, the server boxes, etc). so I'm happy to do what I can to help you debug this problem (collect packets, send logs, etc) if it manifests for us but not for you. Just let me know.

-- dNb

[jjzhuang]

no internal iphone/ipod users have reported excessive battery drain so far.

i've been using iphone sync since April and only saw excessive drain in early iphone early beta.

please keep in mind that in terms of push we don't do anything special for iphone. All other devices use the same push mechanism.

one thing i did notice is that iphone seems to get stuck on 300 second push cycle and not exploring larger push window as other devices does. for example, from our server log i can see some windows mobile devices eventually settle on close to 3000 second push cycle. push window is adaptive and client is supposed to explore the optimal window size by try-and-error. somehow internal iphones don't seem to adapt. even my own iphone is simply doing 300 cycle.

That said, however, even at a short cycle like 300 seconds, my phone has no problem holding battery whatsoever. Battery usually lasts close to 24 hours on a single charge. I have to say that I don't listen to music or watch video on it. I only sync email/contacts/calendar, plus browse internet.

[dNb]

Thanks for the followup, that is comforting. I'm guessing you guys have a range of models for testing as well (orig, touch and 3G). We'll let you know what we find soon.

-- dNb

[jjzhuang]

yes we have all 3 types. not too surprisingly in terms of battery life 3G is actually the worst among the 3.

btw, I do know some devices, not iphone, on at&t network had battery issue. that's because we users buy other gsm devices and insert SIM cards from at&t, they may not be using the proper plan. as a result, they are often on a network that doesn't support long open connections.

i have personally observed that at&t network close idle tcp connection at precisely 60 seconds. this causes the push client to churn every 60 seconds. that caused excessive battery drain on those devices, and the only cure was to turn off push.

but on iphone i have not seen this behavior, since most likely at&t iphone plan makes sure the network doesn't close long idle connections for iphones.

there's actually a very simple thing to do. anyone has this issue should collect sync.log and file a support ticket. we can easily see whether there's anything unusually from the log. otherwise simply speculating on forum doesn't get us anywhere. thanks!

[dNb]

Originally Posted by jjzhuang

but on iphone i have not seen this behavior, since most likely at&t iphone plan makes sure the network doesn't close long idle connections for iphones...otherwise simply speculating on forum doesn't get us anywhere. thanks!

But speculating is fun!

Seriously, though, here's an interesting idea that may be worth including in your data collection: AT&T offers a data plan and an Enterprise data plan. I have yet to see any indication by anyone that the two have any technical differences to them but wouldn't it be interesting to find that the latter came with a longer idle connection timeout and hence avoided the same churn you've seen before? Seems highly improbable to me based on everything I've read, but hey.

-- dNb, not speculating. no sir.