EVIndexing issue & Call with Support

[EVGarrison]

Hi,

Since upgrading to EV 11.0.1 we've had to reinstall the Index redistributable 4 times in 2 months. Clearly indicating something wrong.

Having just got off the phone with Support i wanted to sanity check what ive been told as im not convinced.

We have been using EV on a VM for 7 years and its been bullet proof really, we ran EV10.0.1 - 10.0.4 for 2 years on a 4x vCPU 8G RAM and it was fine.

I understand with EV11.0.1 its brought about some great new features and extra bits and therfore more resource intensive but he has told me to give the VM 32 cores and 32G RAM, move the pagefile to a seperate disk also.

Do any of you see any value in this? I do monitor the CPU and RAM daily and the only time it looks stressed is during the mailbox run in the evening, which i would expect.

We have a compellent SAN thats tiered storage (SSD\FC\Iscsi), so i cant see how moving the pagefile is going to have any improvements, 32 cores and then double the ram to 32G i feel like ive been fobbed off.

Anyone got any opinions?

Regards

[WiTSend]

The CPU/RAM combination depends on a lot of factors, but if you're not seeing it hit it's limits then 32/32 sounds a bit excessive. I've run on VM for years and with the proviso that the VM resources be dedicated, not shared, I haven't had problems.

More to the point, specifically what problems are you seeing with the indexes? 

[EVGarrison]

The problem we have seen goes beyond my technical knowledge of EV, but it manifests itself in 2 ways. We suddenly see a splurge or failed indexes in the EVevent log and the Index service fails start. with fairly vague errors\descriptions but the error code points you to the technote that explains how to ReInstall the Indexes.

(delete the EVva in IIS and the APpPool, delete tehe EVIndexes subdir and then reinstall the msi from the media.

This fixes the problem and the system runs fine for a week or 2, maybe 3. But in 2 months we have had to do it 4 times.

We didnt have any such problems in previous versions, most likley to do with the inception of lots of new features\performance\design gains from 11.0.1.

Either way, something isnt right. Hopefully we will get to the bottom of it in time.

[WiTSend]

We've seen some of the same issues, with little to show for fixing them.  We have to do the same things you are doing.  We've disabled antivirus on the servers and increased the maintenance on our SQL servers but continue to experience similar problems with our indexes.  I'll be interested to see any other responses and recommendations.

[AndrewB]

just out of curiosity, how many EV servers do you have, how many mailboxes are you archiving, are you using DA/CA/Clearwell, is SQL on a dedicated server seperate from EV, were all your indexes upgraded to 64bit before you upgraded to EV 11?

for what it's worth, the recommended basic specs are 8 cores and 16GB RAM. in a virtual environment, the page file and moving things to seperate "disks" isnt really applicable.

[warnox]

I came across the same indexing issues recently on 11.0 and after a multitude of problems it was finally fixed by Method 2 from the MS article below (3). Been stable for about 2 months since. The first two Symantec articles were also suggested as possible fixes.

I did first try the re-installation as you say but the service would fail to start again following the next reboot.

1. https://support.symantec.com/en_US/article.TECH204416.html

2. https://support.symantec.com/en_US/article.HOWTO59060.html

3. https://support.microsoft.com/en-us/kb/2803161 (Method 2)

Hope this helps.

[Gill2]

EVGarrison,

I am also experiencing the same. I have reistalled the redistributable.msi 3 - 4 times. https://support.symantec.com/en_US/article.HOWTO59060.html

We are also on EV11.0.1. I was thinking of installing CHF1. Hoping updating will fix this.

We are running our system with 8x vCPU 16G RAM.