Proliant DL380 G3 (or g4, not sure), won't post, dead cache battery, do I need a new module as well?

[XOIIO]

Hey all, my server has complained for a looooong time about the raid/cache battery being bad, however today I restarted, and that was it, I have no post, no video, keyboard becomes unresponsive for capslock and numlock after a short while, though it is fully running, all the fans are going and the hard drives are spinning up.

I **bleep** the cache battery out, and checked the voltage, 1.2 volts, I am guessing it perhaps dropped below the minimum voltage it could function with.

I am hoping that is the issue, anyways I am going to get a replacement, but I have two questions.

1: Do i need to get the cache module as well? I am wodnering if it maybe had some volatile ram or something that could go bad from lack of power? Looking at the module when turned on it had an orange light and a green light (the green being the one closer to the rear of the server), more info on that would be nice

2:  Will my raid configuration be ruined? I have two raids set up with the hard drives, it's no big loss if I have to reinstall, but I would lose a few things.

[waaronb]

A bad cache battery, even on the older Smart Arrays of the G3/G4 series won't cause the server to not boot up.

It will disable the write cache on the controller, but that's about the extent of it.

The cache module itself should be fine... it's actually using DRAM and the battery is supposed to be able to power the memory for up to 3 days. The newer flash-backed cache for newer controllers uses NVRAM and a super-cap to provide just enough power to write the DRAM contents of the cache to NVRAM in the event of a failure... but the older modules were all BBWC, not FBWC. :)

The batteries used are NiMH... there would be 4 of them in series to provide 4.8V. It's entirely possible you'd be measuring 1.2V per battery right after they get a charge, but really these battery packs only have a production life of 4-5 years before they start giving charge errors.

So, besides replacing the battery pack which is good, there's your issue with the server not POSTing, which is bad...

You could start out by removing all add-in cards and all memory except the bare minimum. Then you can see if it POSTs okay. If so, start adding pieces back in until you hopefully figure out which one is the problem.

Problems where the server won't even POST can be tricky to troubleshoot because you don't always have the same diag lights and indicators... you might still hear it beeping though, and the pattern of beeps can give you some clues.

There are also some LEDs on the system board itself which can sometimes be useful in interpreting where the problem is. The troubleshooting manuals for Proliant servers are somewhere on the HP site so a search should find them.

Once you get it going, your RAID config should be okay. Even if the systems NVRAM is reset or the RAID controller forgets it's settings, each drive has information on itself and it will inform the controller if needed.

It's the same fun magic that lets you move drives from one system into another one and have it recognize the drives and boot up.

[XOIIO]

Alright, not sure why the forum censored something in my original post lol.

It is really strange, I just uninstalled RRAS and rebooted and then nothing happened, I just have a wireless card in it right now, when I have time I'll head back and poke around more with a troubleshooting manual, though there weren't any obvious error lights. I'm also going to bring a netbook to test out the monitor just in case, though it seems to be fine.

I'm pretty sure the battery from this is the factory one, I got this server when it was being upgraded by the university, actually combined two servers so that I have 6 G B of amand two 2.4ghz dual core xeons. It's really nice, I'm just hoping this isn't a serious issue.

Is there a BIOS battery anywhere on it? Didn't spot one, I imagine there is a DIP switch since it's a server. Guess I'll find out more when I check it with a manual.

I also read somewhere that it can be the riser board causing an issue? Seems unlikely though, this thing was purring like a kitten until today.

[XOIIO · Accepted solution]

Err, alright then.

Well, it's a G4 not a G3, and, it's working again *shrugs*. Went there, set up the netbook to find out what the codes meant, turned it on and it works again no problem. No clue what gives but that is certainly nice.