[MoritzF]

Thanks for your answer.

We actually have MCS 2.3.0.3. installed, and on the MiSeq there doesn't seem to be an update available. May I ask which version you're running?

Ever since the previous failed run with memory issues I had decided that we reboot the MiSeq before every run we start, because even at the time I found it suspicious that overclustering alone would cause a system crash. But appearantly, it did not prevent the crash this time.

[mcnelson.phd]

MCS is now at v2.4. If you don't see an upgrade button on the MiSeq itself, I'd suggest downloading the installer directly from the Illumina support site and installing the update. It's only been out for about two weeks now, and the upgrade button ironically didn't show up till after our FAS installed the upgrade the day after it was released (he was here checking the alignment of the optical table so he installed the upgrade as well).

I'd still suggest having your FAS come out and do a more thorough diagnostic to make sure the single board computer doesn't need to be replaced. That would also likely include more RAM since I believe they're now including 16GB or RAM instead of the 12GB that was initially included.

Edit: I just checked our system, our version of MCS is 2.4. For some reason only 2.3 is listed on the Illumina website. Not sure why that is. I'll further note that the upgrade button for MCS didn't appear till after our FAS installed the 2.4 upgrade, so it might not actually be being pushed out even though our FAS and someone in San Diego told me that the update was live as of March 24th.

[MoritzF]

That now is interesting.

I am a bit angry anyway how people manage to install subpar computers into a machine that is about 180 000 $ worth. Initially it contained an i5 (or was it i3?) that needed to be replaced later. Now you're telling me they are increasing the RAM afterwards.
I would think that a with a machine soo expensive they would find a way not to save a few hundred $ by installing cheap versions of somewhat important parts. How much is 4Gb RAM these days??? But then again, I'm not working in industry. But it doesn't make me happy.

[mcnelson.phd]

To be somewhat fair to Illumina, back when they started producing the MiSeq, they weren't producing nearly the same volume of data that they are now. Even three years ago when we got our system, I doubt they had a very good idea of what kind of data load these v3 600 cycle kits would create even though they knew the v2 upgrades were coming.

Contact Illumina support to see what's going on with availability of the 2.4 MCS upgrade, and as I've said before, get your FAS out to look things over.

[GenoMax]

There is a bulletin available on iCom about the 2.4 upgrade that has the following:

A bug has been found in the newly released MiSeq Control Software v2.4 (MCS) software. The bug makes MCS unable to read SmallRNA and TruSeq Custom Amplicon (TSCA) sample sheets properly. If you attempt to begin a run using either of these workflows, MCS throws an error when the sample sheet is loaded.

As a result, Illumina seems to have pulled the 2.4 upgrade from the site. A 2.4.1 would be eventually released to address the problem above.

We have installed MCS 2.4 (above problem does not affect us) and it does speed up v.3 runs significantly.

[MoritzF]

Yes, I already contacted them yesterday, but since I saw this thread I figured why not ask you guys?

I personally guess that when they delivered the MiSeq they already knew or at least considered that someday they would produce an update in chemistry/machine parts that would produce a lot more data. And I think that even with v2 cartridges the RAM was not exactly too much.
My point is that a really good computer and especially RAM can be bought for the price of a sequencing kit (cartridge plus flowcell).

Put another way: If you were given the decision (at the time you bought your MiSeq) to either buy it as is or buy the same machine but with a much better computer and more RAM than you might probably need at the moment just to be sure it runs smoothly for the extra cost of, say, 2000 $, what would you have chosen?
Saving on peanut parts when selling a diamond machine just leaves a bad taste for me, that's all.

[mcnelson.phd]

Ah, that makes sense now, and that bug doesn't affect us at all either so that's probably why our FAS never mentioned it to me since he's usually pretty good at notifying us of these types of things.

And yes, the 2.4.x upgrades really speed up v3 runs. We did a 150 cycle run immediately after the upgrade and I thought I screwed up the sample sheet when it finished in about 12 hrs.

[GenoMax]

@MoritzF: Idea of over-engineering a piece of equipment (in anticipation that it will last well past its design life) seems to have gone out of the window with the focus on relentless development and replacement cycles in today's economy.

[kcchan]

That now is interesting.

I am a bit angry anyway how people manage to install subpar computers into a machine that is about 180 000 $ worth. Initially it contained an i5 (or was it i3?) that needed to be replaced later. Now you're telling me they are increasing the RAM afterwards.
I would think that a with a machine soo expensive they would find a way not to save a few hundred $ by installing cheap versions of somewhat important parts. How much is 4Gb RAM these days??? But then again, I'm not working in industry. But it doesn't make me happy.

I don't understand what you are angry about. Illumina upgraded every MiSeq they sold to include an i7 CPU and 16Gb of RAM as part of the V2 package. Everyone got the upgrades so nobody is losing out. On the other hand I don't think a car manufacturer would upgrade your car if their latest engine had an extra 100HP.

[GenoMax]

There is a bulletin available on iCom about the 2.4 upgrade that has the following:

As a result, Illumina seems to have pulled the 2.4 upgrade from the site. A 2.4.1 would be eventually released to address the problem above.

MCS 2.4.1 is now available on iCom. It may be prudent to wait for a week in case your MiSeq is used for a critical application to avoid problems with any other unforeseen bugs.

[cement_head]

Just installed 2.4.1 - will report back if there's issues.

[DJ Flowcell]

As another data point, we have the latest MCS software 2.4 and have been experiencing crashes still as well. We were offered the same solution , "power cycle, power cycle, power cycle" - seems to be the go-to for tech help... also, I would advise making sure RTA is closed in the task manager because there can be a sync issue with the software and hardware during runs which will render the data useless.

[GenoMax]

As another data point, we have the latest MCS software 2.4 and have been experiencing crashes still as well.

That is odd. We have multiple MiSeq and crashes have never been a problem. Our MiSeq are on a private network.

How are you running your MiSeq's? Is your MiSeq on a public network (without an anti-virus program/SCEP )? Do you apply windows updates?