I've had it, I Google Polyserve and all I see are stories about how great it is and HP bought it. Well, this blog is for those of us to deal with it day in and day out and hate it with a passion. It seems to cause more downtime than in prevents, messes with SANs in odd ways, and is a pain to backup up.
If any other loathers of Polyserve find this, please, please post. Any if any fans find it, I'd also like you to help convince the rest of us why this is worth it. I know that many a manager thinks it's the greatest thing ever (on paper), so I'm more than happy to have people post autonomously. It's easy to get a free gmail account, so go do so and smile at work without anyone knowing who you are.
6 comments:
I spelled 'anonymously' incorrectly. Sorry about that. I was caught up in the moment.
I find it hard to believe that no one else has posted here. I guess all other companies are smart and ditch this worthless product. Bless you for creating this, it has brightened our IT dept.'s dismal day due to polyserve... and the fact that Management still thinks it has value.
I have been forced into using this Craptastic piece of software as well. And I have to say, finding information on it has been a pain.
Also, Backing it up is something else... Our CIO bought in to DataProtector and Polyserve on HP EVA disk....
First things first, we use Polyserve on windows nodes just serving CIFS shares. Backing it up directly gives us thousands of file locks, which are a result of windows granted, however, NTFS has the shadow copy so open file backups is possible with additional licenses from any commercial backup software. But not when its on polyserve. Polyserve does not support shadow copys so open file backups are not possible. In order to do backups we are supposed to snapshot the volumes in order to get a backup of these. The volumes are very large and always changing which makes these snapshots eat up LOTS of space on the VERY expensive HP EVA system (3k/disk@300GB for FC and 500GB for what they call FATA)
also they claim this to be a "clustering" solution, yet, if/when a node fails, all connections are killed as the virtual IP is moved to another node. Making the failure very apparent and causing file corruption issues. Polyserve engineers have said that doesn't mean it went "down", even during an upgrade to 3.6 it brought all connections down to ALL nodes to join the last node into the cluster but that didn't count as an outage because the "systems were up" just inaccessible...
In sort... PolyServe IS JUNK
lol, nice mitch... I think we should make T-Shirts
Holy Crap... Maybe a t-shirt is the best way to tell my company that their software sucks.
We use PolyServe and I agree on all points of it being CrapTastic. I really don't see a single advantage of this over Sun Cluster or Veritas Cluster... the concurrent write just causes more problems then benefit and any decent cluster file system can come close to the performance using nice interconnects.
The only reason we use it is because it's "cheap" and because Veritas didn't install the cluster properly during eval time and the failover took too long.
I'm no Graphic Artist so sign me up if you ever design that t-shirt.
I guess its not so bad after all ? Half a dozen complaints...
We have been using it for 6 months or more (on linux) and it works as advertised. We did a fair bit of testing before it went live and we now have several terra-bytes of disk served via nfs and a few days ago was tested when a sys admin inadvertently shut down one of the nodes. All services continued to run none of the applications even noticed.
Post a Comment