[DRBD-user] Identifying High iowait on DRBD Computers

Robinson, Eric eric.robinson at psmnv.com
Fri Feb 18 01:52:51 CET 2011

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


Yvette,

> > Here's what I know so far: the iowait problem happens when the DRBD 
> > nodes are connected, even when they are both UpToDate.
> 
> Eric, you stated in an earlier email that this began 
> happening some weeks ago, and before that, things were fine.
> 

I did? I don't recall saying that. As far as I know, the problem started
fairly suddenly yesterday or the day before.


> have you taken a cursory look at the operating environment 
> (dasd space utilization, new application(s) being added or 
> changed, new yum/apt-get upgrades (and product installs), etc?
> 

It's an FTP server. No users log into it except through FTP. The RAID 5
array utilization goes up steadily in small increments as new files are
added, but I can't think of anything that would have caused this. No new
upgrades or apps have been installed. Also, this behavior happended to
BOTH cluster nodes, the first one yesterday (and it had to be rebooted)
and now the second one today (and it had to be rebooted). Now the drbd
primary is back to the first one.

> when i'm stumped technically (which happens way too often!) i 
> take a look at what else is going on with the environment, 
> and it's amazing what one can find (a user:  "oh, right, we 
> forgot to mention that app that was changed to read in the 
> world and sort it", you know...)

Fair enough.

> 
> if one of your apps / processes changed processing 
> characteristics, or if one or more files grew too much, or 
> ...  many things such as these could cause your system's 
> "performance profile" to change.
> 

At my desk, I watch sar output all day long. The cluster was fine for a
year, now it is misbehaving. I reallu cannot think of anything that
changed except the volume of data.

> i'm not defending drbd nor am i critical of it, and from all 
> the emails i've read, your latest test seems to be in the 
> right direction.  isolate the variables is usually the 
> quickest route to solving performance issues.
> 

I'm a drbd fan. I have several drbd clusters and this is the first one
that has given me much trouble.

--Eric






Disclaimer - February 17, 2011 
This email and any files transmitted with it are confidential and intended solely for yvette at dbtgroup.com,drbd-user at lists.linbit.com. If you are not the named addressee you should not disseminate, distribute, copy or alter this email. Any views or opinions presented in this email are solely those of the author and might not represent those of Physicians' Managed Care or Physician Select Management. Warning: Although Physicians' Managed Care or Physician Select Management has taken reasonable precautions to ensure no viruses are present in this email, the company cannot accept responsibility for any loss or damage arising from the use of this email or attachments. 
This disclaimer was added by Policy Patrol: http://www.policypatrol.com/



More information about the drbd-user mailing list